Petrogenesis of Extracaldera Rhyolites at Yellowstone Volcanic Field: Videncee for an Evolving Silicic Magma System North of Yellowstone Caldera

UNLV Retrospective Theses & Dissertations

1-1-2005

Petrogenesis of extracaldera rhyolites at Yellowstone volcanic field: videnceE for an evolving silicic magma system north of Yellowstone Caldera

Nicole Marie Nastanski University of Nevada, Las Vegas

Follow this and additional works at: https://digitalscholarship.unlv.edu/rtds

Repository Citation Nastanski, Nicole Marie, "Petrogenesis of extracaldera rhyolites at Yellowstone volcanic field: videnceE for an evolving silicic magma system north of Yellowstone Caldera" (2005). UNLV Retrospective Theses & Dissertations. 1846. http://dx.doi.org/10.25669/mkjk-5nhm

This Thesis is protected by copyright and/or related rights. It has been brought to you by Digital Scholarship@UNLV with permission from the rights-holder(s). You are free to use this Thesis in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s) directly, unless additional rights are indicated by a Creative Commons license in the record and/ or on the work itself.

This Thesis has been accepted for inclusion in UNLV Retrospective Theses & Dissertations by an authorized administrator of Digital Scholarship@UNLV. For more information, please contact [email protected]. PETROGENESIS OF EXTRACALDERA RHYOLITES AT YELLOWSTONE

VOLCANIC FIELD: EVIDENCE FOR AN EVOLVING SILICIC MAGMA

SYSTEM NORTH OF YELLOWSTONE CALDERA

Nicole Marie Nastanski

Bachelor of Science University of Pittsburgh 2002

A thesis submitted in partial fulfillment of the requirements for the

Master of Science Degree in Geoscience Department of Geoscience College of Sciences

Graduate College University of Nevada, Las Vegas August 2005

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 1429720

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

UMI

UMI Microform 1429720

unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Thesis Approval ITNTV The Graduate College University of Nevada, Las Vegas

MARCH, 28 -.20. 0 5

The Thesis prepared by

NICOLE NASTANSKI

E n titled

PETROGENESIS OF EXTRACALDERA RHYOLITES AT YELLOWSTONE VOLCANIC

FIELD : EVIDENCE FOR AN EVOLVING S IL IC IC MAGMA SYSTEM NORTH OF

YELLOWSTONE CALDERA

is approved in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Examination Committee Chair

Dean of the Graduate College

Examtndtion Commitvee Member

m ittee M em berExaminatior

Graduate College Faculty Representative

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT

Petrogenesis of Extracaldera Rhyolites at Yellowstone Volcanic Field: Evidence for An Evolving Silicic Magma System North of Yellowstone Caldera

Nicole M. Nastanski

Dr. Terry L. Spell, Examination Committee Chair Associate Professor of Geology University of Nevada, Las Vegas

Rhyolites erupted at 526 - 80 ka north of Yellowstone Caldera may record the most

recent migration of the Yellowstone melting anomaly, yet their relation to the magma

system that produced caldera-forming eruptions at 2.1, 1.3, and 0.64 Ma was previously

unclear. New pétrographie, geochronology, and geochemical data show that 10 of 12

rhyolites define a spatially (Norris-Mammoth corridor), temporally (326 - 80 ka), and

chemically related group derived from a single magma system independent of the

caldera-related magma system. These data are consistent with the establishment of a

magma system of substantial longevity (>200 ka) and long magma residence times (up to

-100 ka). This system is characterized by significant early magma mingling events

followed by chemical evolution consistent with fraetional crystallization, and subsequent

rejuvenation of residual crystal mush. Silicic magmatism beneath the Norris-Mammoth

corridor may represent the onset of a new volcanic cycle, which identifies this area as a

potential location of future caldera-forming eruptions at Yellowstone.

Ill

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS

ABSTRACT...... iii

LIST OF TABLES...... vi

LIST OF FIGURES...... vii

ACKNOWLEDGMENTS...... viii

CHAPTER 1 INTRODUCTION...... 1 The Geologic Problem...... 2 Research Objectives...... 3

CHAPTER 2 GEOLOGIC SETTING ...... 8 Stratigraphy of the Yellowstone Plateau...... 9 The Third Volcanic Cycle ...... 10 The Plateau Rhyolite...... 11 Geology of the Norris-Mammoth Study Area...... 12 Tectonic Controls on Voleanism ...... 14

CHAPTER 3 PREVIOUS WORK...... 20 Geochronology...... 20 Geochemical Data...... 22

CHAPTER 4 SAMPLE STRATEGY AND ANALYTICAL METHODS...... 24 Sample Collection ...... 24 Sample Preparation ...... 24 LOI Calculations ...... 25 Petrography...... 26 "'**Ar/^^Ar Geochronology ...... 26 XRF/ICP-MS Geochemistry...... 28 TIMS Isotopic Analyses of Sr, Nd, and Pb ...... 28 Geochemical Modeling ...... 29

CHAPTER 5 PETROGRAPHY...... 32 Obsidian Creek Member Rhyolites ...... 33 Roaring Mountain Member Rhyolites ...... 38

CHAPTER 6 ^®Ar/^Âr GEOCHRONOLOGY ...... 46 ""Âr/^Âr Results...... 46

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 7 GEOCHEMISTRY ...... 62 Major Element Chemistry ...... 62 Trace Element Chemistry ...... 63 Nd, Sr, and Pb Isotope Data ...... 67

CHAPTER 8 PETROGENESIS...... 82 Hybrid Source Models ...... 82 Rayleigh Fractionation Models ...... 84 Mixing Models for the Mingled Lavas ...... 88 ^^Sr/^Sr, versus Epsilon Nd Mixing Models ...... 91 Pb Isotope Mixing Models ...... 92

CHAPTER 9 DISCUSSION...... 115 Origin of the Extracaldera Rhyolites ...... 115 Evolution of the Extracaldera Rhyolites ...... 119 Models for Silicic Voleanism ...... 130 Implications for a d*** Volcanic Cycle ...... 139 Summary and Conclusions ...... 143

APPENDIX A '^'^Ar/^^Ar Data Treatment and Analytical Data...... 150

APPENDIX B Major and Trace Element Geochemistry Data ...... 172

APPENDIX C Sample Outliers...... 177

REFERENCES...... 183

VITA...... 192

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES

Table 2.01 Stratigraphy of the Yellowstone Plateau volcanic field ...... 16 Table 2.02 General Stratigraphy of intracaldera and extracaldera members of the Plateau Rhyolite...... 17 Table 4.01 Sample locations ...... 31 Table 5.01 Point count results of selected Obsidian Creek member samples ...... 40 Table 5.02 Point count results of mafic enclaves from the Obsidian Creek mingled lavas...... 41 Table 5.03 Point count results of selected Roaring Mountain member samples ...... 42 Table 6.01 Results of "**^Ar/^^Ar laser fusion and step heating analyses with previous K/Ar ages for comparison...... 51 Table 7.01 Sr, Nd, and Pb isotope data ...... 70 Table 8.01 Nd and Sr hybrid source models ...... 93 Table 8.02 Partition coefficients used in Rayleigh fractionation modeling ...... 94

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES

Figure 1.01 Loeation of Yellowstone National Park ...... 7 Figure 2.01 The Yellowstone melting anomaly track...... 18 Figure 2.02 Map of Norris-Mammoth field area...... 19 Figure 3.01 Nd and Sr isotopic ratios of Yellowstone basalts and rhyolites ...... 23 Figure 5.01 Photomicrograph of mafic-silicic interface of Appolinaris Spring mingled lava...... 43 Figure 5.02 Photomicrograph of reaction-rimmed clinopyroxene ...... 43 Figure 5.03 Photomierograph of mafie-silieic interface of Gardner River mingled lava...... 44 Figure 5.04 Photomicrograph of resorbed sanidine ...... 44 Figure 5.05 Photomicrograph of mafic-silicic interface of the Grizzly Lake mingled lava...... 45 Figure 6.01 "‘Âr/^Âr isochron for sample OC-5 ...... 52 Figure 6.02 '’'Âr/^Âr age spectrum for sample OC-5...... 52 Figure 6.03 isochron for sample GR-2 ...... 53 Figure 6.04 '’Âr/^Âr age spectrum for sample GR-2...... 53 Figure 6.05 "^*Âr/^Âr isochron for sample CC-7...... 54 Figure 6.06 '"Âr/^Âr age spectrum for sample CC-7...... 55 Figure 6.07 '^'Âr/^Âr isoehron for sample CC-7 ...... 55 Figure 6.08 '’“Ar/^Âr age spectrum for sample CC-7...... 56 Figure 6.09 '’Âr/^Âr isochron for sample CC-7 ...... 56 Figure 6.10 "^Âr/^Âr isochron for sample RF-2 ...... 57 Figure 6.11 '^'Âr/^Âr isochron for sample GH-2 ...... 58 Figure 6.12 '^'Âr/^Âr isochron for sample PH-5 ...... 58 Figure 6.13 isochron for sample LD-1 ...... 59 Figure 6.14 isoehron for sample GLM-2 ...... 59 Figure 6.15 "^°Ar/^Âr isochron for sample GRM-1 ...... 60 Figure 6.16 "'‘Âr/^Âr isochron for sample AS-1 ...... 60 Figure 6.17 ‘^'^kxl^Âx isochron for sample WP-1 ...... 61 Figure 7.01 Lebas classification diagram ...... 71 Figure 7.02 Harker variation diagrams...... 72 Figure 7.03 Harker variation diagrams (eont.)...... 73 Figure 7.04 Age versus trace element plots...... 74 Figure 7.05 Nb versus trace element plots...... 75 Figure 7.06 REE plot of Obsidian Creek rhyolites ...... 76 Figure 7.07 REE plots of Roaring Mountain rhyolites ...... 77 Figure 7.08 REE plots of mingled lavas...... 78 Figure 7.09 ^’Sr/^^Sp versus SNd plot of extracaldera rhyolites and mingled lavas 79 Figure 7.10 ^^Sr/^^Sr, versus age plot...... 80

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 7.11 versus W°^Pb and ^"^Pb/^^b plots...... 81 Figure 8.01 ^^Sr/^^Sp versus SNd hybrid source models...... 95 Figure 8.02 Nb versus La and Eu FXL models of Obsidian Creek rhyolites ...... 96 Figure 8.03 Nb versus Th and Ta FXL models of Obsidian Creek rhyolites ...... 97 Figure 8.04 Nb versus La, Sm and Eu FXL models of Roaring Mtn. rhyolites ...... 98 Figure 8.05 Nb versus Yb, Hf, and Ta FXL models of Roaring Mtn. rhyolites ...... 99 Figure 8.06 Nb versus Rb and Ba FXL models of Roaring Mtn. rhyolites ...... 100 Figure 8.07 Major element mixing models for GRM-la and GLM-2 ...... 101 Figure 8.08 Major element mixing models for GRM-la and GLM-2 (eont.) ...... 102 Figure 8.09 Major element mixing models for A S-la...... 103 Figure 8.10 Major element mixing models for AS-la (eont.) ...... 104 Figure 8.11 GIB trace element plot and mixing model for GRM-la ...... 105 Figure 8.12 GIB trace element plot and mixing model for GLM-2 ...... 106 Figure 8.13 GIB trace element plot and mixing model for AS-la ...... 107 Figure 8.14 REE plot and mixing model for AS-2a ...... 108 Figure 8.15 ^’Sr/^^Sp versus SNd mixing models for GRM-la and GLM-2 ...... 109 Figure 8.16 ^^Sr/^Sp versus SNd mixing models for A S -la...... 110 Figure 8.17 ^°®Pb/^*^'^Pb versus ^°^Pb/^*^ Pb mixing models for GRM-la and GLM-2 ...... [...... I ll Figure 8.18 ^^^Pb/^'^Pb versus ^^^Pb/^^'^Pb mixing models for GRM-la and GLM-2 ...... 112 Figure 8.19 ^°^Pb/^°'^Pb versus ^^^Pb/^'^'^Pb mixing models for A S-la...... 113 Figure 8.20 ^®^Pb/^‘^'’Pb versus ^®*Pb/^'’'*Pb mixing models for A S-la...... 114 Figure 9.01 Eruptive periodicity of the Gbsidian Creek and Roaring Mountain rhyolites...... 146 Figure 9.02 Huppert and Sparks (1988) model for silicic voleanism...... 147 Figure 9.03 Bachmann and Bergantz (2004) model for silicic voleanism ...... 148 Figure 9.04 Multistage model for extraealdera rhyolite voleanism...... 149

V lll

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS

This project was funded by the Nevada Isotope Geochronology Laboratory,

University of Nevada, Las Vegas; the Bemada E. French Foundation, University of

Nevada, Las Vegas; the Graduate and Professional Student Association, University of

Nevada, Las Vegas; and the Geological Society of America.

I would like to thank Terry Spell, my advisor and committee chairperson, for his

support and guidance with this project, for numerous helpful conversations, and for

making this project a priority of his over the past few years. Also a special thank you to

my other committee members. Gene Smith, Rod Metcalf, and Steve DeBelle for giving

their time and support throughout this project. Thanks Gene for your close involvement

with many aspects of this project, and for your encouragement and advice regarding my

future endeavors. Thanks Rod for your ever-lasting enthusiasm for igneous petrology,

especially regarding our work in Yellowstone.

A special thank you to Kathy Zanetti for her invaluable help with interpreting data

and also for her support and assistance while working in the NIGL lab. And finally, thank

you to my friends and family for their support and patience.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 1

INTRODUCTION

An ongoing debate in igneous petrology concerns the origin of voluminous high-

silica rhyolites and focuses on whether large volumes of silicic magmas are stored as

pluton-size stable systems on timescales approaching I Ma (Smith, 1979; Hildreth, 1981;

Halliday et al., 1989; Davies et al., 1994; Reid et al., 1997; Heumann et al., 2002), or

alternatively, are generated and erupted rapidly on timescales of « 100 ka (Sparks et al.,

1990; Mahood, 1990; Reid and Coath, 2000; Bindeman and Valley, 2001; Bachmann et

al., 2002). The Yellowstone Plateau Volcanic Field (YPVF), a major Quaternary caldera-

forming silicic magma system in North America (Figure 1.01), has produced voluminous

eruptions of high silica rhyolite throughout its three volcanic cycles. Each cycle

culminated in the formation of a large caldera (Figure 1.01) at ~ 2.1, 1.3, and 0.64 Ma

(Lanphere, et. al, 2002) establishing a recurring explosive eruptive history. The YPVF is

thus an ideal location to study silicic voleanism.

Post-collapse voleanism of the third volcanic cycle (<0.64 Ma) has produced a group

of basalts and two petrographically distinct groups of extraealdera rhyolites (Obsidian

Creek and Roaring Mountain members) that include mingled lavas. These high silica

rhyolite domes/flows erupted in the Norris-Mammoth corridor (Figure LOI and 2.02) and

thus, mark the most recent evolution of the magmatic system at the leading edge of the

northeast migrating Yellowstone melting anomaly. However, their eruptive age,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. petrogenesis, and genetic relationship to associated young basalts and concurrently

erupted intracaldera rhyolites is incompletely understood. The extraealdera rhyolites

therefore offer an opportunity to study the most recent development of a major caldera-

forming silicic magma system and a potential site of a future 4* caldera cycle.

The Geologic Problem

The Yellowstone Plateau volcanic field has produced the most voluminous and

repeated large-scale eruptions of the three major Quaternary, caldera-forming silicic

magma systems in North America (Yellowstone Plateau volcanic field, Wyoming; Long

Valley caldera system, California; the Valles Caldera system. New Mexico). Of the

current and ongoing work in Yellowstone, none has focused specifically on the young

extraealdera rhyolites, yet they potentially offer the opportunity to study the initial stages

of development of a caldera-forming magma system and the role of basalt in generating

and modifying silicic magmas. Study of these rhyolites will also allow models for

production and storage of rhyolite magmas in the upper crust to be tested.

Current models (Hildreth et al., 1991; Christiansen, 2001; Vazquez and Reid, 2002)

suggest that the extraealdera rhyolites are the product of small, independent magma

batches produced from basalt induced partial melting of crustal rocks. The possibility that

the extraealdera rhyolites were derived from a long-lived evolving magma system similar

to the one associated with the voluminous intracaldera rhyolites (Christiansen, 1984;

Hildreth et al., 1991; Vasquez and Reid, 2002) was dismissed due to their small volume,

wide distribution, and range of ages (Hildreth et al., 1991). The petrography,

geochronology, and geochemical/isotopic characteristics of these rhyolites, however are

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. poorly documented and incomplete. Their origin, evolution, and timing of eruption are

therefore incompletely understood.

Re-examination of previously published trace element data (Hildreth et. al, 1991)

shows that the extraealdera rhyolites may alternatively be related to a single evolving

silicic magma system. Additionally, Vasquez and Reid (2002) have found that magma

storage timescales for one extraealdera rhyolite (Gibbon River flow at 90 ka; Obradovich,

1992) range up to 170 ka and average 40 ka based on model ages obtained from U-Th

isotope results for zircons. If the extraealdera rhyolites were produced from an evolving

magma system of substantial size and/or longevity, then they may represent the initial

stages in the development of a fourth caldera-forming cycle. Their location in the Norris-

Mammoth corridor (Figure 1.01) at the leading edge of the Yellowstone melting anomaly

may indicate a migration of the current sub-caldera magma system, or development of a

new magma system north of Yellowstone Caldera.

This study proposes to test the hypothesis of Hildreth et al. (1991) that the

extraealdera rhyolites represent a series of isolated, independent magma batches that are

short-lived and chemically distinct versus a single long-lived magma system that evolved

over time. Detailed petrography, "*®Ar/^^Ar dating, ^^*^Th/^^*U and zircon

dating, and geochemical/isotopic compositions were used to place constraints on the

origin, evolution, magma storage timescales, and eruptive age of the extraealdera

rhyolites.

Research Objectives

The following are the main research objectives of this study.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1) Identify potential source rocks o f the extraealdera rhyolites.

S t, Nd, and Pb isotopic compositions yield constraints on the crustal and mantle

sources for these magmas, and were used to test the argument from previous work

(Hildreth et al., 1991) that the extraealdera rhyolites are not related to the main silicic

magma system beneath the caldera complex.

2) Document the eruptive chronology o f the extraealdera rhyolites.

"*°Ar/^^Ar dating on all twelve domes/flows established reliable age eonstraints on

extraealdera rhyolite voleanism and provided the chronologic framework for interpreting

geochemical and petrogenetic relationships among the extraealdera rhyolites.

3) Determine petrogenetic relationships between the extraealdera rhyolites and the

voluminous intracaldera rhyolites.

Previous work (Hildreth et al., 1991; Vazquez and Reid, 2002) indieates that the

Yellowstone extraealdera rhyolites are unrelated to rhyolites erupted from the main

subcaldera reservoir. However, re-examination of these data has shown that the

extraealdera rhyolites may have been produced from an evolving magma system similar

to subcaldera magma system, which produced the partly eontemporaneous intracaldera

rhyolites (Chapter 3). The extraealdera rhyolite magma system(s), however, appears to be

distinet from the main sub-caldera system based on previous trace element and isotopic

compositions. A complete geochemical/isotopie data set obtained in this study was used

to test these results.

4) Determine petrogenetic relationships among the Obsidian Creek and Roaring

Mountain member extraealdera rhyolites.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The extraealdera rhyolites were originally assigned to two members, a porphyritic

Obsidian Creek member and a predominantly aphyric Roaring Mountain member, based

on their stratigraphy and petrography (Christiansen and Blank, 1972) (Chapter 2).

Detailed petrography and geoehemieal/ isotopic compositions were used to constrain

petrogenetic relationships among the rhyolites, and geochemical modeling using these

data sets contributed to understanding their evolution. More specifically, these data have

constrained whether the two extraealdera rhyolite members represent multiple ehemieally

distinct magma batches or were erupted from a single magma system.

5) Assess the petrogenetic links between the contemporaneously erupted extraealdera

basalts and rhyolites.

The elose spatial relations of the extraealdera rhyolites and basalts within the Norris-

Mammoth corridor along with the oecurrence of mingled lavas of mafic and silicic

composition suggests that a genetic relationship exists. The mingled lavas from the

Obsidian Creek Member (Chapter 2) are significant to this study because they document

the co-existence of basaltic and rhyolitic magma. The previously reported K/Ar dates

(Obradovich, 1992) of the extraealdera rhyolites and basalts indicate that they erupted

contemporaneously. These dates suggest that eruption of basalt may have ceased at -200

ka, whereas rhyolite continued to erupt until -80 ka. Termination of basaltic voleanism

with continued rhyolitic voleanism may imply the establishment of a silicic magma body

large enough to block intruding basaltic magma and prohibit it from reaching the surface

(Doe, et al., 1982; Huppert and Sparks, 1988; Annen and Sparks, 2002). Models were

tested in order to understand the role of extraealdera basalts in the genesis/modification

of extraealdera rhyolite magmas.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6) Establish magma residence timescales.

The extraealdera rhyolite domes and flows have average volumes o f - <1 km^ and

total only - 7 km^ which is dramatically less than the voluminous intracaldera rhyolites

(Hildreth et al., 1991). These small volumes may imply the rapid generation of silicic

magma batches that rise to the upper crust, erupt, and crystallize. However, recent ^^^Th/

ion microprobe data (Vasquez and Reid, 2002) show that zircons range up to 170 ka

older than the eruption age establishing substantial magma residence times, which

suggests a larger, long-lived magma system. These data combined with new Th/ U

and ^^^Pb/^^^U zircon data from several additional extraealdera rhyolites (Spell and

Nastanski, 2004) further constrained magma residence timescales and contributed to

testing models regarding the genesis, evolution, and longevity of the extraealdera magma

system(s).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Yellowstone National Park Montana Mammoth

Yellowstone National Study Site Park

Idaho Norris Wyoming

Calderas I - III

A N 0 km 50 km

Figure 1.01. Insert shows loeation of Yellowstone National Park in western North America. To the right is the volcanic field defined by three caldera-forming cycles: I (unnamed) at 2.1 Ma, II (Island Park/ Henry's Fork ealdera) at 1.3 Ma, and III (Yellowstone ealdera) at 0.64 Ma. Norris-Mammoth study area highlighted by black box. Modified from Christiansen (2001).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 2

GEOLOGIC SETTING

Explosive silicic voleanism at Yellowstone clearly marks the most recent location of

a melting anomaly, whereas older silicic volcanic centers southwest of the volcanic field

mark its origin and migration. Pierce and Morgan (1992) and Perkins and Nash (2002)

propose that voleanism along the Yellowstone melting anomaly originated along the

Oregon-Nevada border at -16.6 Ma as the voluminous Columbia Plateau flood basalt

eruptions were nearing a peak. However, Christiansen and Yeats (1992) noted that a clear

linear progression did not begin until -12.5 Ma. Perkins and Nash (2002) show that the

melting anomaly is best defined by the ignimbrites produced from successive explosive

volcanic centers that become younger to the northeast. Figure 2.01 shows the northeast

propagation of the late Cenozoic major volcanic centers that record the southwestward

movement of the North American Plate over the Yellowstone melting anomaly. Plume

versus nonplume models for the origin of the Yellowstone hotspot as well as its timing of

formation have been a topic of ongoing debate (Anders and Sleep, 1992; Geist and

Richards, 1993; Humphreys and Dueker, 1994; Parsons et al., 1994; Camp, 1995;

Humphreys et al., 2000; Christiansen et al., 2001; DePaolo and Manga, 2003), but are not

the focus of this study.

Regardless of the opposing models, the Yellowstone Plateau formed from a thermal

melting anomaly and is characterized by northeast propagating ealdera complexes and a

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. southwest-widening topographic swell thought to result from a plume head flattening

beneath the southwest moving lithosphere (Anders and Sleep, 1992). The three caldera-

forming events at Yellowstone mark the location of a persistent subcaldera magma

reservoir during the past 2 Ma.

Stratigraphy of the Yellowstone Plateau Volcanic Field

The eruptive history of the volcanic field is defined by three cycles that produced

major caldera-forming eruptions at 2.1, 1.3, and 0.64 Ma (Lanphere, et. al, 2002) (Figure

1.01). The associated ignimbrites produced are the Fluckleberry Ridge, Mesa Falls, and

Lava Creek tuffs, respectively. These ignimbrites are formations assigned to the

Yellowstone Group (Christiansen 2001) and collectively their volume accounts for more

than half of the total (~ 6,500 km^) erupted material at Yellowstone. Minor events that

accompanied the three volcanic cycles are eruption of precaldera and postcaldera

rhyolites and marginal basalts. Table 2.01 summarizes the stratigraphy of the three

volcanic cycles as originally described by Christiansen and Blank (1972), and later

modified by Christiansen (2001). The stratigraphie relations of the precaldera and

postcaldera rhyolites and basalts are constrained by K/Ar dating (Obradovich, 1992),

dating (Gansecki et al., 1996), and paleomagnetic studies (Christiansen, 2001).

The ages of caldera-forming ignimbrites have been constrained by recent ''^Ar/^^Ar dating

(Lanphere et al., 2002).

Volcanic rocks of the third cycle (Table 2.01) are the best preserved and have

therefore served as a model to understand the development of the previous two cycles.

The most recent volcanic cycle, which began at 1.1 Ma (K/Ar) and peaked at 0.64 Ma

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ( ^Ar/ At) with eruption of more than 1000 km^ of the Lava Creek Tuff and subsequent

formation of the Yellowstone Caldera, is also the most significant to this study.

The Third Volcanic Cycle

A typieal caldera-forming event at Yellowstone is considered to be part of a

“Resurgent Cauldron Cycle” as originally described by Smith and Bailey (1968). Each

cycle includes the following sequential stages: (1) “A period of gentle uplift of an area

slightly more extensive than the source of magma, which results in radial fractures

commonly accompanied by eruption of rhyolite lavas and minor basalt flows that erupt at

the margins of the volcanic system” (Christiansen 2001, p. 115). Stage one is represented

in the third volcanie eycle by eruption of the precaldera Mount Jackson and Lewis

Canyon rhyolites (Table 2.01) beginning at 1.2 Ma (K/Ar) from radial fractures that

formed during the initial development of the volcanic cycle. The Shotgun Valley, Warm

River, and Undine Falls basalts (Table 2.01) erupted during this time at the margins of

the future Yellowstone ealdera. (2) “Caldera-forming pyroclastic eruptions in which large

volumes of silicic magma are emplaced as ignimbrites. Partial evaeuation of the shallow

magma chamber results in collapse of the magma chamber roof, forming a ealdera”

(Christiansen 2001, p. 115). Stage two is represented by eruption of the Lava Creek Tuff

(Tahle 2.01) at 0.64 Ma followed by collapse of the magma chamber roof forming the

Yellowstone Caldera, an 85 km long by 45 km wide depression (Christiansen, 2001). (3)

“Post-collapse voleanism during which resurgent doming and silicic voleanism is

controlled by ring fractures and tectonic faults” (Christiansen 2001, p. 115). Stage three

is represented by postcollapse rhyolites named the Plateau Rhyolite (Tables 2.01 and

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.02), which are described below in detail. Although not apparent in the first two cycles,

post-collapse voleanism of the third cycle does include resurgent doming within the

ealdera (Christiansen, 2001).

The Plateau Rhyolite

The Plateau Rhyolite is divided into five members, which are separated into an

older post-resurgence intracaldera rhyolite group (Upper Basin Member), and a younger

rhyolite group of intracaldera members (Mallard Lake and Central Plateau) and

extraealdera members (Obsidian Creek and Roaring Mountain) (Table 2.02). Voleanism

producing the Plateau Rhyolite began with the older group at 0.64 Ma and continued until

at least 70 ka with eruption of the Pitchstone Plateau (Central Plateau member) based on

K/Ar dating (Obradovich, 1991). Marginal basalts contemporaneous with the Plateau

Rhyolite are listed in Table 2.01.

The Mallard Lake member is comprised of a single flow that erupted prior to

formation of the Mallard Lake dome, whereas the younger Central Plateau member

intracaldera flows erupted after its formation. The voluminous Central Plateau flows are

commonly referred to as the intracaldera rhyolites, which occur within the basin of

Yellowstone ealdera with the exception of several that spill over the ealdera wall near the

Norris Geyser Basin. The intracaldera rhyolites are comprised of twenty-one massive

flows that have thicknesses, which exceed 300 m (Christiansen, 2001). Previously

reported K/Ar ages (Obradovich, 1992) indicate the intracaldera rhyolites erupted from

-164 - 70 ka, contemporaneous with the Obsidian Creek and Roaring Mountain member

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. extraealdera rhyolites. The Obsidian Creek and Roaring Mountain member rhyolites are

described in detail below in a discussion of the study area.

Geology of the Norris-Mammoth Study Area

With the exception of two rhyolite units (Cougar Creek and Riverside flows, Roaring

Mountain Member) that erupted -25 km to the west, all domes/flows of the Obsidian

Creek and Roaring Mountain members are located along a 22 km stretch within the

Norris-Mammoth corridor (Figure 2.02). The corridor is associated with hydrothermal

activity related to north striking extensional faults, and is also the location of several

extraealdera basalt vents (Bennett, in progress). The close spatial relations of the

extraealdera rhyolites and basalts within the Norris-Mammoth corridor along with the

occurrence of mingled lavas of mafic and silicic composition document the spatial and

temporal relationship that exists between the rhyolitic and basaltic magmas of this

bimodal system.

The extraealdera rhyolites are divided into two members based on their stratigraphie

position and a distinct petrography. Christiansen and Blank (1972) originally described

the Obsidian Creek member as porphyritic rhyolite domes that include two mingled

rhyolite-basalt lava flow complexes. The Roaring Mountain member is characterized by

aphyric rhyolite.

Obsidian Creek Member

The Obsidian Creek member as originally defined includes the Gibbon Hill,

Paintpot Hill, Willow Park, Apollinaris Spring, and Landmark domes (Figure 2.02),

which range in diameter from 0.5 - 1.5 km. The largest of these domes (Gibbon Hill)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rises more than 250 m above its base. Rhyolite of all domes contains abundant

phenocrysts of quartz, sanidine, and plagioclase (Christiansen, 2001). The Willow Park

and Gibbon Hill domes were dated by the K/Ar method at 316 ka and 116 ka

respectively, indicating that the Obsidian Creek Member rhyolites were erupted over at

least a 200 ka interval (Obradovich, 1992).

Both the Gardner River and Grizzly Lake mingled lavas have been described by

Iddings (1899), and the Gardner River lava has been studied by Fenner (1938; 1944),

Wilcox (1944), Hawkes (1945), and Struhsacker (1978). Fenner (1944) concluded that

Gardner River rhyolite erupted into deeply eroded older basalt and incorporated clasts of

it in the rhyolite. Wilcox (1944) was the first to map and study the petrography of the

Gardner River mingled lava and concluded that the rhyolite and basalt components must

have both been at least partly liquid at the time of formation based on intrusive relations

and the fact that quartz and sanidine xenocrysts were present in the basalt. Hawkes

(1945) agreed with Wilcox (1944) that the two components were commingled as liquids,

but believed that the complex originated from a composite feeder intrusion that had

basaltic margins and a rhyolitic core. Struhsacker (1978) also proposed that the Gardner

River lava was formed from the intimate commingling of rhyolitic and basaltic magmas.

Roaring Mountain Member

The Roaring Mountain member as originally defined includes the Crystal Spring,

Gibbon River, Obsidian Cliff, and Riverside flows, and the Cougar Creek dome (Figure

2.02). All are rhyolite flows, petrographically distinguished from the Obsidian Creek

Member by being aphyric (Christiansen, 2001). All but one of the Roaring Mountain

member rhyolites were dated by the K/Ar method at 400 - 80 ka (Obradovich, 1991).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Tectonic Controls on Voleanism

The following discussion is based on Christiansen (2001) who summarized the late

Cenozoic tectonic history of the region surrounding the Yellowstone Plateau volcanic

field through the work of Pardee (1950), Myers and Hamilton (1964), Fraser et al. (1964),

Ruppel (1972), Love et al. (1973), Love and Keefer (1975), Smith et al. (1977), Anders et

al. (1989), Pierce and Morgan (1992), and Smith and Braile (1994).

The Yellowstone Plateau volcanic field and the eastern Snake River Plain are flanked

by north-trending mountain ranges that represent tilted fault blocks bounded primarily by

normal faults. These extensionally driven topographic features are characteristic of the

Basin and Range tectonic province south of the Snake River Plain. Basin and Range

extension is believed to have initiated in the northern Great Basin Province during the

mid-Miocene, but exactly when in the Miocene tectonic extension began in the

Yellowstone region has been debated.

Previous investigators have recognized a broad area of late Cenozoic major fault

zones that trend south from the northern Yellowstone Plateau and end abruptly at the

ealdera margin. The location of volcanic vents of the extraealdera rhyolites, mingled

lavas, and various vents of the Swan Lake Flat basalt appear to be largely controlled by

one of these linear normal fault zones, which defines the Norris-Mammoth corridor

(Figure 2.02). This trend of volcanic vents eontinues across the Yellowstone ealdera

through the intracaldera Central Plateau rhyolites to the Mariposa Lake basalt flow

approximately 20 km outside the southern ealdera margin (Christiansen, 2001). The

continuation of these features along a major fault zone emphasizes the structural control

of voleanism in the Yellowstone Plateau. In addition, the fact that the three ealdera

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. collapse events all occurred at the intersection of these major fault zones with the

volcanic axis of the Snake River Plain-Yellowstone Plateau shows that regional as well

as local structures have acted to focus the location of voleanism.

Hydrothermal activity at Yellowstone is also largely tectonically controlled. The

Norris-Mammoth corridor is notable for its past and present hydrothermal activity outside

of the ealdera. The presence of faults and fractures in this area allows meteoric water to

circulate and be heated in the subsurface by a source believed to be magmatic (Fournier

and Pitt, 1985; Fournier, 1989). The 43 km length of this area shows a variety of

hydrothermal features that include boiling springs and geysers at Norris, fumarolie

activity and acid alteration at Roaring Mountain, and moderate temperature travertine

deposits at Mammoth.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■DCD O Q. C g Q.

■D CD

C/) C/)

8 Table 2.01 Stratigraphy of the Yellowstone Plateau volcanic field. From Christiansen, 2001. (O' Volcanic Precaldera Caldera-forming Ignimbrite Postcaldera Contemporaneous Age Cycle Rhyolite (Yellowstone Group) Rhyolite Marginal Basalts Basalt of Snake River Osprey 3. Madison River 3" CD Plateau Basalt of Geode Creek ■DCD Rhyolite Swan Lake Flat O Q. Third Gerrit C a Falls River o 3 as Pleistocene Basalt of Maiposa Lake "O o Lava Creek Tuff Mount Jackson Undine Falls CD Q. Lewis Canyon Basalt of Warm River Basalt of Shotgun Valley Island Park Basalt of Narrows ■D CD Second Mesa Falls Tuff Big Bend C/) C/) Big Bend Ridge Pliocene First Huckleberry Ridge Tuff Snake River Junction Butte ■DCD O Q. C 8 Q.

■D CD

C/) C/)

8 ci'

3 3" Table 2.02. General stratigraphy of intracaldera and extracaldera members of the Plateau Rhyolite. CD From Christiansen, 2001. ■DCD O Q. C Plateau Rhyolite Age Group Caldera Event Intracaldera Flows Extracaldera Flows a O Central Plateau 3 ■D Younger postcollapse Member Obsidian Roaring O Late doming Creek Mountain

CD Mallard Lake Member Member Member Q. Upper Basin Member Older post resurgence Early resurgent doming ■D CD

C/) C/) 50 N 120 W n o w 50 N _

0 250 500 km

Yellowstone Melting Anomaly Track

16.4 Ma - 40 N ' t BJ HR OH

Colorado

Plateau

Rio Grand Rift

120 W n ow

Figure 2.01. The Yellowstone melting anomaly track and associated silicic volcanic centers. Modified from Perkins and Nash (2002). The melting anomaly track is believed to have originated at -16.6 Ma along the Oregon-Nevada border. Its migration is tracked by the location of silicic volcanic centers shown in black. HR = High Rock caldera complex (16.4 Ma); McD = McDermitt volcanic field (vf); OH = Owyhee-Humbolt vf (15.2 Ma); BJ = Bruneau- Jarbridge vf (12.7 Ma); TF - Twin Falls vf (10.5 Ma); P = Picabo vf (10 Ma); H = Heise vf (6.6 Ma); YP = Yellowstone Plateau vf (2.1 Ma). 87sr/86sr isotope boundary (dashed line) marks eastern limit for Mesozoic plutonic rocks with inital < 0.706. Higher ^^Sr/^^gj- values occur to the east of the 0.706 line.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. • Mammoth" v

Norris ,

Figure 2.02. Map of Norris-Mammoth field area north of Yellowstone Caldera (black dash). Obsidian Creek rhyolites (dark pink); PH - Paintpot Hill dome, GH = Gibbon Hill dome, LD = Landmark dome, AS = Appolinaris Spring dome, WP - Willow Park dome, GRM = Gardner River mingled lava, GLM = Grizzly Lake mingled lava. Roaring Mtn. rhyolites (medium pink): RF = Riverside flow, GR = Gibbon River flow, OC = Obsidian Cliff flow, CS = Crystal Spring flow, CC = Cougar Creek dome. Associated basalts (yellow): mr = Madison River, slf = Swan Lake Flat. Stratigraphically related Lava Creek Tuff (purple) and intracaldera rhyolites of the Central Plateau member (light pink). Selected major normal faults shown in black (ball on downthrown side of normal fault). GIS and geology data from the National Park Service.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 3

PREVIOUS WORK

There has been no previous work focusing specifically on the petrogenesis of the

extracaldera rhyolites, however several studies have generated data that contribute to

constraining the generation, storage, and evolution of these rhyolite domes/flows. Models

to explain the petrogenesis of these silicic magmas have been proposed based on

preliminary K/Ar geochronology and geochemical/isotopic measurements.

K/Ar ages obtained by Obradovich (1992) on several Obsidian Creek and Roaring

Mountain member rhyolites provide initial constraints on the timing of extraealdera

rhyolite and basalt volcanism, and Vasquez and Reid (2002) provide data regarding

magma residence timescales for one extraealdera unit (Gibbon River flow). Doe et al.

(1982) and Hildreth et al. (1984, 1991) have determined major and trace element

chemistry and Sr, Nd, and Pb isotopic data on several extraealdera rhyolites. Several

investigators (Friedman et al., 1974; Hildreth et al., 1984; Bindeman and Valley, 2000,

2001; Bindeman et al., 2001) have added to the isotopie dataset by analyzing several

extraealdera rhyolites for oxygen isotopes.

Geochronology

Sanidine from one (Cougar Creek) and obsidian from three (Crystal Spring, Obsidian

Cliff, and Cougar Creek) of five Roaring Mountain member rhyolites, and sanidine from

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. two (Willow Park and Gibbon Hill) of seven Obsidian Creek member rhyolites were

analyzed by the K/Ar method. Results from these analyses (Table 6.01) place the eruptive

ages of the extraealdera rhyolites at -400 - 80 ka (4 of 6 fall within a range o f-180 ka -

80 ka) making them contemporaneous with the intracaldera Central Plateau member,

erupted between -165 - 70 ka (Obradovieh, 1992). Eruption of the associated

extraealdera basalts occurred between -600 - 200 ka as constrained by stratigraphy

(Christiansen, 2001) and a single K/Ar analysis of the Osprey basalt flow (-220 ka,

Obradovich, 1992). Thus, basaltic volcanism may have ceased north of Yellowstone

Caldera at -200 ka whereas extraealdera rhyolites continued to erupt up until at least 80

ka.

The extraealdera rhyolites at Yellowstone have average volumes of < 1 km^ and total

only - 7 km^, which may imply short residence times if they were erupted from

individual magma batches as is thought (Hildreth et al., 1990). Recent ^^'^Th/^^^U ion

microprobe data from Vasquez and Reid (2002) on the Gibbon River flow (90 ka, K/Ar

eruptive age) however show that zircon ages range up to 180 ka older than the eruption

age, establishing substantial residence timescales similar to those recognized for young

small volume rhyolites at Long Valley, California (Reid et al., 1997). Vasquez and Reid

(2002) have determined zircon saturation temperatures of -800 °C for the Gibbon River

flow based on the experimental constraints of Watson and Harrison (1983), indicating

that the magma system cooled to < 800 °C as long as 270 ka ago.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Geochemical Data

Isotopic data from Hildreth et al. (1991) show that the high-siliea extraealdera

rhyolites cannot be related to the main subcaldera magma system that produced the

voluminous intracaldera rhyolites. Furthermore, they suggest that the small volume, wide

areal distribution, and range of ages of the extraealdera rhyolites support a derivation

from several independent magma batches as opposed to a single evolving magma

reservoir.

The extraealdera rhyolites (7 of 12 analyzed) show lower CNd values and generally

higher ^^Sr/^Sn values than intracaldera rhyolites erupted from the main subealdera

reservoir (Figure 3.01), which supports derivation from separate sources. Freidman et al.

(1974), Hildreth et al. (1984), and B indeman and Valley (2001) have shown that the

extraealdera rhyolites have higher ô'^O values compared to the low- ô'*0 intracaldera

lavas. Bindeman and Valley (2001) show that the intracaldera rhyolites, which erupted

shortly after ealdera collapse, have unusually low 5**0 values (0 to +3%). This is

interpreted to represent remelting of the hydrothermally altered, '*0-depleted base of the

ealdera (Bindeman and Valley, 2001). Extraealdera rhyolites have normal ô’*0 values

(+5 to +8%) indicating they were derived from a relatively uncontaminated magma

supply that is unrelated to the subcaldera magma system.

Hildreth et al. (1991) also use traee element data to support their interpretation that

the extraealdera rhyolites represent a distinet magma system. This eonelusion is based on

the faet that the extraealdera rhyolites are distinet in their traee element abundances from

the intracaldera rhyolites.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 1 1------1— • Basalts 0.5126 ■ Intracaldera Rhyolites 0.5125 A Ignim brites + Extraealdera Rhyolites 0.5124

0.5123 Main 0.5122 Reservoir ENd 143 Nd 0.5121 -10 144 Nd 0.5120

0.5119 -15 0.5118 Extraealdera Rhyolites 0.5117

0.5116

0.5115 _L _L _L 0.705 0.709 0.713 0.717 0.721 0.725 87 Sr/86sn

Figure 3.01. Initial Nd and Sr isotopic ratios of Yellowstone basalts and rhyolites. Extraealdera rhyolites show they are distinct from the ignimbrites and intracaldera rhyolites related to the main magma reservoir due to their lower and slightly higher ^^Sr/^^Srj isotopic ratios. Modified from Christiansen, 2001.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 4

SAMPLE STRATEGY AND ANALYTICAL METHODS

Sample Collection

Fieldwork for this project was completed during June 2003. The sampling strategy

was based on mapping by Christiansen (2001) and consisted of collecting multiple

samples from each Obsidian Creek and Roaring Mountain member domes/flows (12

units). All samples collected from outcrop were trimmed of weathering rinds in the field,

and were later petrographic^lly examined in order to select a subset of representative

samples for "''^Ar/^Ar dating and geochemical/isotopic analysis (Table 4.01). Samples

selected for geochemical/isotopic analysis showed minimal devitrification and displayed

no evidence of alteration. Samples selected for dating contained fresh glass (free of

devitrification or hydration) and/or clear, unaltered sanidine phenocrysts.

It is possible that these units may be compositionally zoned, or that they may

represent more than a single eruptive event, however due to the budget limitations of this

project, a single sample is used to represent the mineralogy/chemistry of each dome/flow.

Sample Preparation

Following sample crushing and sieving, mineral separations of sanidine from

porphyritic rhyolites and glass from aphyric rhyolites were handpicked under a binocular

microscope at the coarsest grain size (0.5 - 2.0 mm) and highest purity possible for

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ''^Ar/^^Ar dating. An appropriate amount of material was picked in order to perform both

laser fusion and furnace step heating analyses if necessary. Sanidine crystals were

washed in an ultrasonic bath of 1 M HF for 3-5 minutes to remove any adhering glass,

and obsidian fragments were washed in an ultrasonic bath of acetone for 3-5 minutes

before being packed for irradiation.

A powder sample from each dome/flow to be used for geochemical and isotopic

analysis was ground from picked rock chips using an agate shatterbox. Approximately 2-

4 g of fine powder were sealed in Pyrex vials and sent to the Washington State University

GeoAnalytical Lab to be analyzed for major and trace elements. Powders were also sent

to the Isotope Geochemistry Laboratory at the University of Kansas for Sr, Nd, and Pb

isotopic analysis.

LOI Calculations

Loss on ignition calculations were conducted to measure the degree of hydration of

the rhyolites. The amount of water and other volatiles retained in a sample is expressed

by the percent loss on ignition (LOI). Table 4 in Appendix B lists the measurements for

each step of the following process. Each ceramic crucible was cleaned and weighed to ±

0.0005 g. Approximately 2.0 - 3.0 g of powdered sample was added to the crucible and it

was weighed again. The original crucible weight was subtraeted from the sample +

crucible weight to calculate the sample weight to ± 0.0005 g. The crucible+ sample was

placed in a 110°C oven for 2 hours and then removed to cool to room temperature. The

weight of the crucible+ sample was measured again to determine the weight loss after

heating to 110°C. Dividing the weight loss after 110°C by the original sample weight and

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. normalizing that value to 100% calculates the LOI for 110°C. LOI values for 1 IO°C

represent the percent H 2 O' lost. Next the crucible + sample were placed in a 950°C oven

for 2 hours and then removed to cool to room temperature. The LOI values for 950°C

were calculated in the same manner as for 110°C, and represent the percent HiO^ lost.

The total loss on ignition for each sample is the combined values for 110°C and 950°C.

Petrography

Fifteen thin sections from Quality Thin section were examined using a pétrographie

microscope to determine phenocryst assemblages and pétrographie textures/

relationships. Thin sections from the Gardner River, Grizzly Lake, Appolinaris Spring,

and Crystal Spring mingled lavas were cut to expose silicic-mafic interfaces in order to

assess the degree of magma mingling/mixing. In addition, thin sections of the mafic

enclaves (GRM-1 a, GLM-2a, and AS-1 a) were cut to obtain representative mineralogy.

Pointcounts of 600 counts/ thin section were conducted to generate representative

proportions of phenocryst phases and groundmass. The modal abundances of individual

phenocryst phases from potential parental magmas were used in Rayleigh fractionation

models (Chapter 8).

"^^Ar/^^Ar Geochronology

'*°Ar/^^Ar dating was performed at the Nevada Isotope Geochronology Laboratory

(NIGL) at the University of Nevada, Las Vegas on representative samples from all twelve

extraealdera domes/flows. CO 2 laser fusion and furnace step heating analyses of sanidine

and obsidian were performed to obtain eruptive ages.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Samples were packed in A1 foil and stacked in a sealed Pyrex tube prior to irradiation.

Samples were irradiated in-core for 1 hour at the McMaster Nuclear Reactor at McMaster

University, Ontario, Canada. During irradiation, significant amounts of Ar isotopes may

form from interfering nuclear reactions with potassium and calcium. To monitor and

correct for reactor-induced interferences, K-glass and CaF] standards were included in the

irradiation package.

“’Âr/^Âr age determinations rely upon the formation of ^Âr from produced

when fast-neutrons bombard samples. Fish Canyon Tuff sanidines (FC-2) with an age of

27.9 Ma (Steven et al., 1967; Cebula et al., 1986) were packed in the tube at 10 mm

intervals to monitor the fast-neutron dose a sample receives. Measured variations in

neutron fluence along the length of the irradiated tube define the J factor. Irradiated

standards (CaFi and K-glass), fluence monitors (FC-2), and samples to be analyzed by

laser fusion were loaded in a Cu tray and placed in a high vacuum extraction line for

analysis using a 20 W CO 2 laser. Laser fusion analysis of CaF 2 standards determined Ca

correction factors of (^W ^A r)ca = 3.97 (± 17.59%) x 10'^ and (^W ^A r)ca = 7.83 (±

8.75%) X 10"^, and analysis of K-glass produced a K correction factor of ('*'’Ar/^®Ar)K=

0.0001 (± 100%). Laser fusion of 3-5 individual FC-2 sanidine erystals monitored a

neutron flux variation of < 2% and determined J factors with a reproducibility of 0.19% -

0.77%.

Samples analyzed by incremental step heating were heated in a double vacuum

resistance furnace. In both methods, gases released during heating were cleaned by GP-

50 SAES getters and then admitted to the mass spectrometer via expansion throughout

the extraction line. Atmospheric argon was measured to determine the mass

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. discrimination of the mass spectrometer. By admitting small volumes of atmospheric

argon from an on-line pipette system to a MAP 215-50 mass spectrometer a '**^Ar/^^Ar

ratio of 289.85 ± 0.13% was determined. This value is lower then the accepted value of

295.5 (Steiger and Jager, 1977), thus indicating that the mass spectrometer discriminates

in favor of ^^Ar relative to "^^Ar. A correction factor of 1.0195 (4 AMU) is therefore

applied to measured isotope ratios. Line blanks averaged 3.0 mV for mass 40 and 0.01

mV for mass 36 for laser fusion analyses and 3.71 mV for mass 40 and 0.01 mV for mass

36 for furnace heating analyses. Data reduction and age calculations were done using

LabSPEC software (B., Idleman, Lehigh University) and Isoplot (Ludwig, 1998).

XRF/ICP-MS Geochemistry

XRF/ICP-MS analyses of major and trace elements were conducted at the

GeoAnalytical Lab at Washington State University. Analyses were performed using a

Rigaku 3370 XRF Spectrometer and a Sciex Elan model 250ICP-MS following the

analytical techniques of Johnson, et al. (1999) and Knaack et al. (1994). Major elements

and Ni, Cr, V, Ga, Cu, and Zu were analyzed by XRF. All other trace elements were

analyzed by ICP-MS. The accuracy and precision of XRF/ICP-MS analyses are reported

in Appendix B.

TIMS Isotopic Analyses of Sr, Nd, and Pb

Whole-rock TIMS analyses for ^’Sr/*®Sr, *^^Nd/'^^Nd, and ^°^Pb/^°^Pb, ^"W^°^Pb,

and ^“^Pb/^'^'^Pb isotopic compositions were conducted at the Isotope Geochemistry

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Laboratory at the University of Kansas using a VG sector 54 mass spectrometer.

Analytical procedures follow Krogh (1973, 1982) and Patchett and Ruiz (1987).

Since the Obsidian Creek and Roaring Mountain member rhyolites contain high

concentrations of Rb (average ~ 220 ppm) and low concentrations of Sr (average ~ 12

ppm) as determined in whole-rock ICP-MS analyses, it is important to correct for the

amount of ^’Sr that accumulated in the rock from decay of ^^Rb since the time of

eruption. Initial *’Sr/*^Sr values at the time of eruption were therefore caleulated for all

rhyolites (including mingled lava components) following the procedure outlined by Faure

(1986). Initial ^^Sr/^Sr values for the mafic components of mingled lavas only differ

from the measured values in the sixth decimal place indicating that they are insensitive to

post-eruptive decay of *^Rb to ^^Sr (Table 7.01).

Geochemical Modeling

Fractional crystallization models were calculated using the Rayleigh fractionation

equation in a Microsoft Excel spreadsheet that utilized bulk distribution coefficients

(Table 8.01), whole-rock trace element concentrations, and modal phenocryst abundances

of parental magmas as input values. Excel spreadsheets were also set up to calculate

major element mixing models for the Obsidian Creek mingled lavas using the mixing

equation of Langmuir et al. (1978). Trace element mixing models were calculated using

IGPET 2000 for Windows by utilizing whole-rock trace element concentrations of end-

member magmas. Isotopic mixing models were calculated using a variation of the

DePaolo mixing equation (1981) in a Turbo Basic program, which calculates isotopic

mixtures given input values of ^^Sr/^Sr, epsilon Nd, ^‘^^Pb/^‘^'*Pb, ^®^Pb/^*^'*Pb, or

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 208pb/204pb isotopic ratios as well as their respective Sr, Nd, and Pb whole-rock element

concentrations. Isotopie mixing ean also be done in IGPET 2000.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■DCD O Q. C g Q.

■ D CD

C/) C/)

8 ci'

Table 4.01. Sample locations Latitude Longitude Unit Sample (UTM) (UTM) Sample Location 3 3" Roaring Mountain Member CD Crystal Spring flow CS-l N 523473 W 4967170 From north tip of flow about 2.25 km east of southern Willow Park. ■DCD Obsidian Cliff flow OC-5 N 522919 W 4961092 From 0.4 km northeast of Lake of the Woods. O Q. Gibbon River flow GR-2 N 520444 W 4944895 From east side of Gibbon River Canyon about 4 km south of Artists Paintpots trailhead. C a Cougar Creek dome CC-7 N 502137 W 4947998 From south side of dome about 2.75 km from the Gneiss Creek trailhead. 3O Riverside flow RF-2 N 495113 W 4945687 Across Madison River from end of service road ~ 1 km from West Entrance. "O Obsidian Creek Member O Gibbon Hill dome GH-2 N 523697 W 4949697 From 0.1 km south of dome summit. Paintpot Hill dome PH-5 N 520972 W 4948196 0.4 km southwest of the Artists Paintpots trail terminus on western slope of dome. CD Q. Landmark dome LD-1 N 523237 W 4961438 From west side of dome about 0.8 km northeast of Lake of the Woods. Grizzly Lake flow GLM-2 N 518960 W 4962360 From southern lobe of flow about 2 km from the south Mount Holmes trailhead. Gardiner River flow GRM-1 N 521412 W 4970070 From east side of Grand Loop Rd. ~ 0.25 km south of Sheepeater Cliff picnic area. ■D Appolinaris Spring dome AS-1 N 525530 W 4967760 From 0.3 km northwest of vent about 4.75 km east of central Willow Park. CD Willow Park dome WP-1 N 525400 W 4969415 From 0.2 km west of vent about 4 km east of northern Willow Park.

C/) C/) CHAPTER 5

PETROGRAPHY

Pétrographie results indicate that the majority of Obsidian Creek domes/lavas are

porphyritic whereas the Roaring Mountain units are sparsely porphyritic or aphyric

rhyolites. Magma mingling is evident in the Gardner River, Grizzly Lake, and

Appolinaris Spring flows of the Obsidian Creek member, as well as in the Crystal Spring

flow of the Roaring Mountain member. Point count results of selected Obsidian Creek

member rhyolites (includes dacite from the Grizzly Lake mingled lava). Obsidian Creek

member mafic enclaves, and Roaring Mountain member rhyolites (all listed in order of

increasing age) are presented in Tables 5.01, 5.02, and 5.03, respectively. Individual

phenocrysts are presented as a percent of the whole rock and as their modal abundance

relative to all phenocrysts. Mafic clots within the silicic hosts (GLM-2, GRM-1, and AS-

1; Table 5.01 and CS-la; Table 5.03) of mingled lavas were point counted and their

abundances are reported separately from phenocryst/modal abundances. The term “clot”

used in the following pétrographie descriptions indicates small (<1-2 mm) clumps of

mafic material within a silicic host, whereas “enclave” refers to cm-sized or larger mafic

material.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Obsidian Creek Member Rhyolites

The Obsidian Creek member rhyolites include the Gibbon Hill (GH-2), Paintpot Hill

(PH-5), Landmark (LD-1), Willow Park (WP-1), and Appolinaris Spring (AS-1) domes

and the Gardner River (GRM-1) and Grizzly Lake (dacite, GLM-2) flows. Included in

this member are three mingled lavas (Appolinaris Spring, Gardner River, and Grizzly

Lake) that contain mafic enclaves (AS-la, GRM-la, and GLM-2a).

Porphyritic Domes

The GH-2, PH-5, LD-1, and WP-1 samples contain 7.5 - 18.1 % phenocrysts set in a

highly devitrified groundmass (Table 5.01). Spherulites are comprised of

microphenocrysts of quartz and feldspar and are surrounded by glass. Phenocrysts are an

assemblage of medium grained (1.0 - 5.0 mm) sanidine -i- quartz ± plagioclase ±

clinopyroxene ± zircon ± oxides. Sanidine occurs as 1.0 - 4.0 mm euhedral to subhedral

blocky grains which are commonly fractured and contain glass inclusions near cores.

Sanidines rarely display Carlsbad twins, but often is un-twinned. Quartz is present as 1.0

- 3.0 mm anhedral, subrounded grains that are commonly fractured. Quartz in the Gibbon

Hill dome is resorbed. Plagioclase is less common and occurs in glomerocrysts, or as 1.0

- 2.0 mm blocky subhedral phenocrysts with albite-law twinning. Clinopyroxene is only

present in the Willow Park dome as 0.5 - 1.0 mm subhedral to anhedral phenocrysts.

Mingled Lavas

The silicic host lavas and associated mafic enclaves (AS-la, GRM-1 a, and GLM-2a)

exhibit varying evidence of physical mixing. Outcrop scale observations indicate that the

relative silicic-mafic proportions of the mingled lavas vary for each flow. Approximately

20% of the Appolinaris Spring mingled lava consists of small mafic enclaves (-1-10 cm)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that have crenulated margins. In the Gardner River mingled lava, -25% silicic lava is

chaotically interlayered with -75% mafic lava disaggregated into 1-20 cm-sized mafic

enclaves. In the Grizzly Lake mingled lava -40% silicic lava is chaotically interlayered

with -60% mafic lava disaggregated into 1 - 10 cm enclaves with crenulated margins.

The Appolinaris Spring and Gardner River silicic host lavas are high-silica rhyolite,

whereas the Grizzly Lake silicic host lava is dacite (Figure 7.01). Associated mafic

components AS-1 a and GRM-la are basaltic-andesite and GLM-2a is an andésite.

The rhyolite component (AS-1) of the Appolinaris Spring mingled lava contains

-15% phenocrysts of sanidine + quartz + clinopyroxene + plagioclase set in a devitrified

spherulitic groundmass that contains opaque oxides and microphenocrysts of

clinopyroxene (Table 5.01). 7.6% of AS-1 comprises sub-spherical (-2 mm diameter)

basaltic-andesite clots with crenulated margins. Crenulated interfaces show

disaggregation of mafic clots and clinopyroxenes that migrate into the rhyolite (Figure

5.01). Sanidine occurs as 1.0 - 2.5 mm euhedral to subhedral blocky phenocrysts

occasionally zoned and embayed. Sanidine is commonly mantled by reaction rims that

occasionally engulf the entire crystal. Quartz is present as 0.25 - 1.0 mm anhedral

subrounded grains. Clinopyroxene is less common and often resides in close proximity to

mafic enclaves. Clinopyroxene is present as 0.25 - 1.0 mm euhedral to subhedral

phenocrysts that are in places mantled by dark reaction rims (Figure 5.02). Plagioclase is

found strictly near mafic clots as small laths < 1.0 mm.

An Appolinaris Spring mafic enclave (AS-la) contains 10.2% phenocrysts of

plagioclase + clinopyroxene + olivine + orthopyroxene -+- quartz set in a holocrystalline

groundmass of plagioclase, clinopyroxene, and opaque oxides (Table 5.02). Plagioclase

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. is present as 0.5 - 1.0 mm albite-twinned laths, but also as large (1.0 - 4.0 mm) sieve

textured, blocky grains. Clinopyroxene occurs as 0.25 - 0.5 mm euhedral to anhedral

phenocrysts. Olivine occurs as 0.25 - 0.5 mm anhedral fractured phenocrysts.

Othropyroxene is present as ~ 0.5 mm euhedral to subhedral laths. Quartz within the

mafic material is present as 0.25 - 0.5 mm anhedral crystals that commonly contain

reaction rims of pyroxene microphenocrysts (Figure 5.01).

Mingling in the Gardner River flow appears to be highly chaotic in the field as well as

on a microscopic scale. The silicic-mafic interface in thin section GRM-1 (Figure 5.03)

has an intricate flame-like crenulation indicating that both components were liquid at the

time of interaction. Many phenocrysts near the silicic-mafic interface exhibit a resorbed

texture indicating disequilibrium. The GRM-1 rhyolite host contains 2.5% phenocrysts of

sanidine + quartz set in a glass groundmass that comprises 15.3% of this thin-section

(Table 5.01). The mafic enclave accounts for the remaining 82.2% of this thin-section,

however the reported percentages are not an accurate representation of the Gardner River

flow since this thin section was preferentially cut to show the mafic-silicic interface. The

rhyolite host contains sanidine that occurs as 0.5 - 2.0 mm subhedral, fractured

phenocrysts with reaction rims and glass inclusions. Many sanidines are resorbed with a

reaction rim (Figure 5.04), or exhibit a sieve texture indicating they underwent extensive

disequilibrium reactions. Quartz in the rhyolitic lava is present as 0.5 -1 .0 mm anhedral,

fractured phenocrysts. Glomerocrysts in which euhedral plagioclase (0.5 - 1.0 mm)

commonly overgrows euhedral to subhedral clinopyroxene (0.5 - 2.5 mm) are also

present within the rhyolitic glass groundmass.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A large (20 cm) mafic enclave (GRM-la) represents the mineralogy for the basaltic-

andesite that mingled with the rhyolite host of the Gardner River flow. The enclave

contains 17.0% phenocrysts of plagioclase + clinopyroxene + orthopyroxene + quartz set

in a holocrystalline groundmass of plagioclase, clinopyroxene, orthopyroxene, and

opaque oxides (Table 5.02). Plagioclase occurs as 0.25 - 3.0 mm euhedral laths and

blocky phenocrysts that are strongly to weakly zoned, or'unzoned. Plagioclase

commonly contains clinopyroxene and orthopyroxene inclusions and exhibits normal to

irregular albite-law twinning. Clinopyroxene and orthopyroxene are present as 0.25 - 1.0

mm euhedral to anhedral phenocrysts that are commonly resorbed. Glomerocrysts in

which euhedral plagioclase (0.5 - 1.0 mm) commonly overgrows euhedral to subhedral

clinopyroxene (0.5 - 2.5 mm) are also present. Quartz xenocrysts occur as 0.5 - 1.0 mm

fractured grains, or as rounded grains with reaction rims.

The dacite component (GLM-2) of the Grizzly Lake mingled lava contains 14%

phenocrysts of plagioclase + quartz + clinopyroxene + orthopyroxene set in an

undevitrified glass groundmass that is flowbanded and includes microphenocrysts of

plagioclase, clinopyroxene, and a few oxides (Table 5.01). Plagioclase is present as

isolated 0.5 - 2.0 mm euhedral laths, and is also present in glomerocrysts as subhedral

blocky grains that often contain clinopyroxene inclusions. Quartz occurs as 0.25 - 1.0

mm anhedral rounded grains. Clinopyroxene is present as 0.25 -1 .0 mm subhedral to

anhedral phenocrysts commonly overgrown by plagioclase. Orthopyroxene is less

common, occurring as 0.25 - 0.5 mm subhedral phenocrysts. The dacite contains mafic

clots (13.5% of the whole rock. Table 5.01) composed of glomerocrysts of plagioclase

(0.5 - 1.5 mm) and clinopyroxene (0.5 - 1.0) set in a holocrystalline groundmass of the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. same minerals. The dacite-mafic clot interfaces are crenulated and are a location where

mafic phase minerals have disaggregated from the clot and migrated into the daeite.

Figure 5.05 shows a mafic clot with abundant disaggregated plagioclase and

clinopyroxene phenocrysts that appear to have migrated into the dacite glass where they

are locally concentrated along flowbanding.

Sample GLM-2a was a large enclave (-30 cm) originally collected because it was

thought to represent the mafic clots/enclaves within the dacite (GLM-2) component of the

Grizzly Lake mingled lava. Petrography however shows that GLM-2a is distinct in its

mineralogy from the abundant mafic clots in GLM-2. GLM-2a contains 24.4 %

phenocrysts of plagioclase + quartz + biotite + sanidine + clinopyroxene + hornblende

(Table 5.02). Plagioclase phenocrysts occur as 1.0 - 2.0 mm subhedral to anhedral,

blocky and tabular grains that are often replaced by calcite, or that are commonly zoned,

and occasionally twinned. Quartz is present as mostly unaltered clusters of interlocking

grains (0.25 - 0.5 mm) that commonly enclose biotite. Less common are larger isolated

quartz xenocrysts (-1 mm) that have oscillatory extinction and are embayed or

commonly rimmed by smaller blocky quartz grains. Biotite is present as 0.25 - 1.0 mm

euhedral to subhedral tabular phenocrysts, unaltered and commonly overgrown by quartz.

Sanidine occurs as 1.0 - 2.0 mm blocky phenocrysts that commonly exhibit oscillatory

extinction, Carlsbad twins and partial calcite replacement. The pyroxenes appear to be

mostly clinopyroxene that occur as < 0.25 mm anhedral, unaltered grains. Hornblende

occur as -1 mm subhedral to anhedral highly altered phenocrysts, many of which exhibit

moderate to complete replacement by calcite. Phenocrysts are set in a holocrystalline

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. groundmass of tabular plagioclase that exhibit a flow alignment around phenocryst

phases.

Roaring Mountain Member Rhyolites

The Roaring Mountain member rhyolites are aphyric and sparsely porphyritic lavas

(Table 5.03), and include the Gibbon River (GR-2) flow, Obsidian Cliff (OC-5) flow.

Crystal Spring (CS-l) flow. Cougar Creek (CC-7) dome, and Riverside (RF-2) flow.

Also included in this member is the mingled lava of the Crystal Spring flow.

Aphyric Lavas

The Obsidian Cliff and Crystal Spring flows are fresh glassy lavas. Sample OC-5

from the Obsidian Cliff flow (Table 5.03) includes trace amounts of plagioclase

microphenocrysts and zircon, and sample CS-l from the Crystal Spring flow (Table 5.02)

includes unidentified microlites with a parallel orientation and micro-glomerocrysts of

pyroxene and plagioclase.

Mingled Lava

The aphyric Crystal Spring lava represents the rhyolite component of newly

discovered mingled lava in the Roaring Mountain member. CS-la is a representative

sample of the mingled lava that shows chaotic inter-layering of the silicic and mafic

components on a microscopic scale. The mineralogy of the mafic clots is described

below, but could not be represented from point count analysis (Table 5.03) because

mingling occurs on too small of a scale. Sixteen percent of sample CS-la is composed of

flattened sub-spherical mafic clots with crenulated margins that are enclosed in a glassy

groundmass with abundant plagioclase and clinopyroxene microphenocrysts and few

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. spherulites. The mafic clots consist of 0.25 -1 .0 mm euhedral, albite-law twinned

plagioclase phenocrysts and 0.25 - 0.5 mm euhedral to anhedral clinopyroxene

phenocrysts enclosed in a fine-grained groundmass.

Porphyritic Lavas

The Cougar Creek dome (CC-7) and Riverside flow (RF-2) are sparsely porphyritic

lavas containing 4.8 - 7.6% phenocrysts, whereas the Gibbon River (GR-2) flow is a

porphyritic lava containing up to 13.1% phenocrysts. Phenocrysts from all flows are

enclosed in a glassy groundmass (Table 5.02). Phenocryst phases are an assemblage of

fine (< 1 mm) to medium (1-5 mm) grained sanidine + quartz ± plagioclase ±

clinopyroxene ± orthopyroxene. Sanidine is present predominantly as 0.5 - 1.5 mm

euhedral to anhedral blocky phenocrysts that commonly exhibit Carlsbad twinning.

Sanidine in the Cougar Creek dome commonly shows slight zoning. Quartz is present as

0.5 - 1.0 mm anhedral subrounded grains in all lavas, and is commonly embayed in the

Cougar Creek dome. Plagioclase is less common and only present in glomerocrysts as

0.5 - 1.0 mm subhedral blocky phenocrysts within the Cougar Creek and Gibbon River

lavas. Clinopyroxene and orthopyroxene phenocrysts (0.25 - 1.5 mm subhedral crystals )

are found as rare glomerocrysts in the Cougar Creek and Gibbon River lavas only.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.01. Point count results of selected Obsidian Creek member samples. Sample GH-2 PH-5 LD-1 GLM-2 GRM-1 AS-1 WP-1 dacite w/ rhyolite rhyolite mafic w/ mafic w/ mafic rhyolite rhyolite rhyolite clots enclave clots rhyolite % Phenocrysts (whole rock) Quartz 1.00 2.66 2.16 2.66 0.83 2.00 2.33 Sanidine 6.33 6.33 8.00 0 1.67 9.00 15.00 Plagioclase 0.16 0 0.16 8.66 0 1.00 0.50 Biotite 0 0 0 0 0 0 0 Clinopyroxene 0 0 0 2.43 0 2.00 0.33 Orthopyroxene 0 0 0 0.33 0 0 0 Olivine 0 0 0 0 0 0 0 Hornblende 0 0 0 0 0 0 0 Accessories 0 1.16 1.66 0 0 1.00 0 Total Phenocrysts 7.49 10.15 11.98 14.08 2.50 15.00 18.16

Modal Abundance Quartz 13.35 26.21 18.03 18.89 33.20 13.33 12.83 Sanidine 84.51 62.36 66.78 0.00 66.80 60.00 82.60 Plagioclase 2.14 0 1.34 61.51 0.00 6.67 2.75 Biotite 0 0 0 0 0 0 0 Clinopyroxene 0 0 0 17.26 0 13.33 1.82 Orthopyroxene 0 0 0 2.34 0 0 0 Olivine 0 0 0 0 0 0 0 Hornblende 0 0 0 0 0 0 0 Accessories 0 11.43 13.86 0 0 6.67 0

% Glass Groundmass (whole rock) Devitrified 92.51 82.00 72.19 0 0 68.33 64.84 Undevitrified 0 7.85 15.83 72.42 15.33 9.01 17.00 Total Groundmass 92.51 89.85 88.02 72.42 15.33 77.34 81.84

% Mafic Clots/Enclave 0 0 0 13.50 82.17 7.66 0 Note: Mafic clots within GLM-2 and AS-1 and the enclave portion of GRM-1 were point-counted, but their mineralogy was not included in % phenocrysts or modal abundances, thus % phenocrysts and modal abundances are representative of the silicic host. Accessory phases include zircon ± oxides

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5.02. Point count results of mafic enclaves from

Sample GLM-2a GRM-la AS-la basaltic- basaltic- andésite andesite andesite % Phenocrysts (whole rock) Quartz 6.66 183 0.17 Sanidine 2H6 0 0 Plagioclase 8J3 12.00 5.17 Biotite 133 0 0 Clinopyroxene 2.16 3.17 2.50 Orthopyroxene 0 1.00 0.50 Olivine 0 0 L83 Hornblende 1.83 0 0 Accessories 0 0 0 Total 24.47 17.00 10.17

Modal Abundance Quartz 27J2 4^8 1.67 Sanidine 8.83 0 0 Plagioclase 34.04 70.59 50.84 Biotite 13.61 0 0 Clinopyroxene 183 18.65 24^8 Orthopyroxene 0 188 4.92 Olivine 0 0 17.99 Hornblende 7.48 0 0 Accessories 0 0 0

% Holocrystalline Groundmass (whole rock) 75.53 83.00 89.83

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sample CS-1 CS-la OC-5 GR-2 CC-7 RF-2 rhyolite w/ mafic rhyolite clots rhyolite rhyolite rhyolite rhyolite % Phenocrysts (whole rock) Quartz 0 0 0 1.83 2.17 2.50 Sanidine 0 0 0 10.33 1.50 5.16 Plagioclase 0 0 0 0.50 0.17 0 Biotite 0 0 0 0 0 0 Clinopyroxene 0 0 0 0.17 1.00 0 Orthopyroxene 0 0 0 0.33 0 0 Olivine 0 0 0 0 0 0 Hornblende 0 0 0 0 0 0 Accessories 0 0 0 0 0 0 Total 0.00 0.00 0.00 13.16 4.84 7.66

Modal Abundance Quartz 0 0 0 13.91 44.83 32.64 Sanidine 0 0 0 78.50 30.99 67.36 Plagioclase 0 0 0 3.80 3.51 0 Biotite 0 0 0 0 0 0 Clinopyroxene 0 0 0 1.29 20.66 0 Orthopyroxene 0 0 0 2.51 0 0 Olivine 0 0 0 0 0 0 Hornblende 0 0 0 0 0 0 Accessories 0 0 0 0 0 0

% Groundmass (whole rock) Devitrified 0 0 0 8&84 0 9234 Undevitrified 100.00 84.00 100.00 0 95.16 0 Total Groundmass 100.00 84.00 100.00 86.84 95.16 92.34

Mafic Clots 0 16.00 0 0 0 0 Note: Sample CS-la represents the silicic-mafic mingled lava of the Crystal Spring flow and does not represent a mafic enclave as in samples GLM-2a, GRM-la, and AS-la from Table 5.02.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.01. Mafic-silicic interface of the Appolinaris Spring mingled lava highlighting the crenulated margin of a representative mafic enclave that disaggregated and the clinopyroxenes have migrated into the silicic host (PPL). An exchanged phenocryst of quartz (right of center) now in the basaltic-andesite enclave is mantled by a reaction rim of clinopyroxene.

Figure 5.02. Example of a clinopyroxene phenocryst mantled by a dark reaction rim in rhyolite of the Appolinaris Spring mingled lava (PPL). Clinopyroxenes within the rhyolite are commonly found in close proximity to basaltic-andesite enclaves and display similar reaction features.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. g # ' 2 mm r Rhyolitic Glass i Si ^ > " '4 '

Figure 5.03. Photomicrograph of the Gardner River mingled lava under PPL showing the crenulated, flame-like margin of a mafic-silicic interface indicating liquid state interaction. The dark material (lower field of view) is altered basaltic-andesite with resorbed phenocrysts of sanidine and pyroxene. The rhyolitic glass (above interface) has a brown margin indicating it was in disequillibrim with the basaltic-andesite it interacted with.

2 mm i

? : Resorbe Sanidine Î

Figure 5.04. Photomicrograph (PPL) of a large (~2 mm) resorbed sanidine phenocryst with a reaction rim at the mafic-silicic margin of the Gardner River flow shows the phenocryst is in disequillibrium with the melt.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.05. Photomicrograph of the Grizzly Lake mingled lava mafic-silicic interface. Top (PPL) and bottom (XPL) show plagioclase (plag) and clinopyroxene (cpx) phenocrysts disaggregating from the mafic enclave and migrating into the dacite where they are concentrated along flowbanding.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 6

^ W ^ A r GEOCHRONOLOGY

"*®Ar/^^Ar Results

Results from laser fusion analyses of single/multiple sanidine crystals and glass

fragments, and furnace step heating experiments are summarized in Table 6.01. The

"'°Ar/^^Ar results from each Roaring Mountain and Obsidian Creek member rhyolite are

presented below in order of increasing age. All uncertainties are reported at 1er. Data

tables that contain the details from analyses and a discussion of data treatment for these

samples can be found in Appendix A.

Roaring Mountain Member

Crystal Spring flow (CS-1)

Ages could not be calculated for this sample since furnace step heating of 42.33 mg

of glass yielded no measurable radiogenic argon ('*‘^Ar*) (Appendix A). The previously

reported K/Ar age of 80 ± 2 ka (Table 6.01) will be used in this study.

Obsidian Cliff flow (OC-5)

Eight laser fusion analyses of multiple (2) glass fragments yielded an average age of

86.8 ± 10.8 ka (Appendix A). Four of eight analyses were statistically excluded to give a

refined population that yielded a mean age of 96.0 ± 6.7 ka and a weighted mean age of

99.1 ± 0.8 ka (Appendix A and Table 6.01). An isochron using four out of eight analyses

defines an age of 81.0 ± 17.5 ka (MSWD = 1.9) with a "**^Ar/^^Ar ratio of 294.0 ±75.0

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Figure 6.01). Although this is a statistically valid isochron, the large uncertainties, low

spread in radiogenic yields, and minimum number of data points (n = 4) makes this a

poorly constrained best-fit line. Therefore, the isochron age is considered unreliable.

Furnace step heating of 38.10 mg of glass yielded apparent ages ranging from ~ 83 ka

to ~ 330 ka (Appendix A). The age spectrum produced is discordant (Figure 6.02).

Although the majority of steps yielded fairly ideal ages (Appendix A), no plateau or

isochron was produced. The weighted total gas age of 105.6 ± 1.0 ka for this sample is

considered the most reliable and is taken as the preferred age (Table 6.01).

Gibbon River flow (GR-2)

Laser fusion analysis of 12 single sanidine crystals yielded a large range of ages from

-102 ka to ~ 200 ka (Appendix A). The 12 analyses visually represent seven distinct age

populations on a cumulative probability diagram (not illustrated). The statistical method

failed for eliminating ages > 2a from the mean age because it reduced the dataset to only

2 analyses. However a valid isochron using 4 out of 12 single crystals yielded an age of

118.0 L 10.0 ka (MSWD = 1.9) with a ''^Ar/^^Ar ratio of336.0± 14.0 (Figure 6.03)

indicating the presence of excess argon. Although the isochron uses the minimum

number of data points, a fairly wide spread of radiogenic yields and low uncertainties

make the best-fit line, and therefore age, well constrained.

Furnace step heating of 36.32 mg of sanidine yielded a discordant age spectrum with

apparent ages ranging from —50 ka to —183 ka (Figure 6.04). No plateau or isochron was

produced and the low gas releases from this analysis (Appendix A) result in an unreliable

weighted total gas age of 133.6±1.0 ka. The isochron age of 118.0 ± 10.0 ka from the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. single sanidine crystal laser fusion analysis is considered more reliable and is taken as the

preferred age (Table 6.01).

Cougar Creek dome (CC-7)

Eleven laser fusion analyses of multiple (2) glass fragments yielded an average age of

371.4 ± 34.0 ka (Appendix A). Four of 11 analyses were statistically excluded to give a

refined population that yielded a mean age of 393.9 ± 17.6 ka, and a weighted mean age

of 399.3 ± 3.6 ka (Appendix A and Table 6.01). An isochron using 6 out of 11 analyses

defines an age 391.0 ± 14.0 ka (MSWD - 2.0) with a '^'^Ar/^Ar ratio of 297.0 ± 5.5

(Figure 6.05). Although this is a statistically valid isochron, five out of six data points

have large uncertainties and have similar radiogenic yields. The isochron is thus

interpreted as being constrained by essentially 2 data points, and so the resulting age and

initial '^'’Ar/^^Ar composition are considered unreliable.

Furnace step heating of 61.52 mg of glass produced a flat concordant age spectrum

excluding a high final age of ~ 599 ka that represents virtually none of the total gas

released (Appendix A). Steps 3 through 9 (65.7 % of the^^Ar released) yield a plateau

age of 411.4 + 1.9 ka (Figure 6.06). This age is slightly younger than the total gas age of

417.0 ± 2.5 ka. A valid isochron is defined by steps 7 through 12 (44.4% of the ^^Ar

released), which cluster/overlap on the isochron (Figure 6.07), and so the age of 378.0 ±

2.0 ka and initial '*°Ar/^^Ar composition of 561.0 ± 120.0 are considered poorly

constrained and unreliable. The plateau age for this analysis is therefore considered more

reliable than the isochron age.

Furnace step heating of 13.82 mg of sanidine produced a flat, nearly concordant age

spectrum (Figure 6.08). The apparent ages range from ~ 317 ka to ~ 395 ka if

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. anomalously young and old ages from the initial and final steps are exeluded (Appendix

A). No plateau was produced, however a total gas age of 368.2 ± 2.0 ka gives a reliable

age. A valid isochron (MSWD = 0.7) is defined by steps 5 through 8 (40.1% of the ^®Ar

released), however the four data points do not represent a large spread in radiogenic

yields (Figure 6.09) and so the age (286.0 ± 28.0 ka) and ""^Ar/^^Ar ratio (373.0 ± 77.0)

are not considered to be well constrained.

Although the furnace step heating analysis of glass produced a plateau age, the

sanidine total gas age of 368.2 ± 2.0 ka is considered to be more reliable and is taken as

the preferred age (Table 6.01). Sanidine analyses yield more reliable ages because

sanidine contains low amounts of initial Ar and holds measurable amounts of potassium

in lattice sites as opposed to glass, which commonly yields anomalous ages as discussed

in Appendix A.

Riverside flow (RF-2)

Fourteen laser fusion analyses of single sanidine crystals yielded an average age of

537.5 ± 15.8 ka (Appendix A). Nine of 14 analyses were statistically excluded to give a

refined population that yielded a mean age of 535.0 ± 4.6 ka and a weighted mean age of

537.5 ± 2.2 ka (Appendix A and Table 6.01). A valid isoehron using 7 out of 14 single

crystal analyses defines an age of 525.8 ± 3.3 ka (MSWD = 1.6) with a "^°Ar/^^Ar ratio of

306.3 ± 4.5 (Figure 6.10) indicating a small amount of excess argon may be present. The

large spread of radiogenic yields for these data points gives a well-constrained and

reliable isochron age, which is taken to be the preferred age (Table 6.01).

Obsidian Creek Member

Gibbon Hill dome (GH-2)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Laser fusion analysis of 13 single sanidine crystals yielded an average age of 185.5 ±

119.4 ka (Appendix A). Five of 13 analyses were statistically excluded to give a refined

population with a mean age of 135.5 ± 4.9 ka and a weighted mean age of 134.3 ± 1.4 ka

(Appendix A and Table 6.01). A valid isochron using 8 out of 13 single crystal analyses

defines an age of 134.3 ± 2.6 ka (MSWD = 1. 6 ) with a "^°Ar/^^Ar ratio of 295.3 ± 1.5

indicating that no excess argon is present. A good spread in the radiogenic yields (Figure

6 .1 1 ) make this a well-constrained and reliable isochron age, which is taken to be the

preferred age (Table 6.01).

Paintpot Hill dome (PH-5)

Fifteen laser fusion analyses of single sanidine crystals yielded an average age of

228.0 ± 17.7 ka (Appendix A). Nine out of 15 analyses were statistically excluded to give

a refined population with a mean age of 214.3 ± 3.0 ka and a weighted mean age of 213.9

± 0.9 ka (Appendix A and Table 6.01). A valid isochron using 8 out of 15 single crystal

analyses defines an age of 208.1 ± 4.9 ka (MSWD = 2.1) with a "*°Ar/^®Ar ratio (305.0 ±

6.0) within uncertainty of atmospheric composition (Figure 6.12). A large spread in

radiogenic yields makes this a well-constrained and reliable isochron age taken to be the

preferred age (Table 6.01).

Landmark dome (LD-1)

Fifteen laser fusion analyses of single sanidine crystals yielded an average age of

219.7 ± 13.1 ka (Appendix A). Six of 15 analyses were statistically excluded to give a

refined population that yielded a mean age of 220.9 ± 3.5 ka and a weighted mean age of

221.9 ± 1.1 ka (Appendix A and Table 6.01). A valid isochron using 11 out of 15 single

crystal analyses defines an age of 226.0 ± 5.5 ka (MSWD = 1.8) with a ^^°Ar/^Ar ratio of

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 288,0 ± 9.5 (Figure 6.13). A good spread in radiogenic yields and a ratio that

indicates no excess argon is present makes this a well-constrained and reliable isochron

age, which is taken to be the preferred age (Table 6.01).

Grizzly Lake mingled lava (GLM-2)

Eight laser fusion analyses of multiple (2-3) sanidine crystals yielded an average age

of 278.1 ± 30.5 ka (Appendix A). Three of 8 analyses were statistically excluded to give

a refined population with a mean age of 266.2 ± 5.2 ka and a weighted mean age of 267.4

± 2.4 ka (Appendix A and Table 6.01). A valid isochron using 5 out of 8 analyses defines

an age of 263.3 ± 3.4 ka (MSWD = 1. 6 ) with a "'°Ar/^^Ar ratio of 299.0 ± 2.5 (Figure

6.14). A large spread in radiogenic yields and a ''°Ar/^^Ar ratio that indicates no excess

argon is present makes this a well-constrained and reliable isochron age, which is taken

as the preferred age (Table 6 .0 1 ).

Gardner River mingled lava (GRM-1)

Twelve laser fusion analyses of multiple (2-3) sanidine crystals yielded an average

age of 299.9 ± 12.6 ka (Appendix A). Four of 12 analyses were excluded to give a refined

population with a mean age of 302.6 ± 3.6 ka and a weighted mean age of 302.0 ± 1.5 ka

(Appendix A and Table 6.01). A valid isochron using 8 out of 12 analyses defines an age

of 300.5 ± 2.8 ka (MSWD = 1.9) with a ''^Ar/^^Ar ratio of 295.9 ± 0.6 (Figure 6.15). A

large spread in radiogenic yields and a "^^Ar/^^Ar ratio that indicates no excess argon is

present makes this a well-constrained and reliable isochron age, which is taken to be the

preferred age (Table 6.01).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Appolinaris Spring dome (AS-1)

Fourteen laser fusion analyses of single sanidine crystals yielded an average age of

312.4 ± 30.0 ka (Appendix A). Six of 12 analyses were excluded to give a refined

population that yielded a mean age of 313.8 ± 4.6 ka and a weighted mean age of 313.6 ±

1.6 ka (Appendix A and Table 6.01). A valid isochron using 9 out of 14 single crystal

analyses defines an age of 316.0 ± 2.2 ka (MSWD = 1.4) with a ‘'^Ar/^^Ar ratio of 281.0 ±

4.8 indicating that no excess argon is present (Figure 6.16). A large spread in radiogenic

yields for these analyses results in a well constrained and reliable isochron age that is

taken as the preferred age (Table 6 .0 1 ).

Willow Park dome (WP-1)

Thirteen laser fusion analyses of single sanidine crystals yielded an average age of

327.3 ± 29.8 ka (Appendix A). Seven of 13 analyses were statistically excluded to give a

refined population with a mean age of 330.8 ± 5.1 ka and a weighted mean age of 330.4 ±

1.8 ka (Appendix A and Table 6.01). A valid isochron using 8 out of 13 single crystal

analyses defines an age of 325.8 ± 2.2 ka (MSWD = 1.4) with a '"^Ar/^^Ar intercept of

301.1 ±1.7 indicating that a small amount of excess argon maybe present (Figure 6.17).

A large spread in radiogenic yields results in a well constrained and reliable isochron age

that is taken as the preferred age (Table 6 .0 1 ).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■DCD O Q. C g Q.

■D CD

C/) 3o' O

(O' Table 6.01. Results of laser fusion and step heating analyses with previous K/Ar ages for comparison. Sample ID K/Ar (ka ± o) '“’Ar/^^Ar (ka ± ct) Laser Fusion '*®Ar/^^Ar (ka ± cr) Furnace Step H eat Mean Weighted Mean Isochron Total Gas Plateau Isochron Roaring Mtn. Member 3. 3" CS-1 80 ±2* NO RESULT NO RESULT NO RESULT NO RESULT NO RESULT NO RESULT CD OC-5 183 ±3 96.0 6 6.7 99.1 6 0.8 81.06 17.5 105.6 6 1.0* (g) no plateau (g) no isochron (g) ■DCD GR-2 90 ±2 142.5 6 0.5 142.5 6 1.4 118.0 6 10.0* 133.6 6 1.0 no plateau no isochron O CC-7 399 ±3 393.9 6 17.6(g) 399.3 6 3.6(g) 391.0 6 14.0 (g) 368.2 6 2.0* no plateau 286.0 6 28.0 Q. C 417.0 6 2.5(g) 411.4 6 1.9 (g) 378.0 6 2.0 (g) a o RF-2 UNDATED 535.0 6 4.6 537.5 6 2.2 525.8 6 3.3* 3 LA ■D Obsidian Creek Member. O GH-2 116±8 135.5 6 4.9 134.3 6 1.4 134.3 6 2.6* PH-5 UNDATED 214.3 6 3.0 213.9 6 0.9 208.1 6 4.9* CD Q. LD-1 UNDATED 220.9 6 3.5 221.96 1.1 226.0 6 5.5* GLM-2 UNDATED 266.2 6 5.2 267.4 6 2.4 263.3 6 3.4* GRM-1 UNDATED 302.6 6 3.6 302.0 6 1.5 300.5 6 2.8* AS-1 UNDATED 313.8 6 4.6 313.6 6 1.6 316.0 6 2.2* ■D CD WP-1 31665 330.8 6 5.1 330.4 6 1.8 325.8 6 2.2*

C/) Note: K/Ar ages from Obradovich (1992). All ''^Ar/^^Ar results are from sanidine unless noted as glass (g). C/) Preferred ages are marked by an asterisk. o c - 5 - glass

0.0036 Isochron age = 81.0 ± 17.5 ka 0.0032 40/36 = 294.0 ± 75.0 MSWD = 1.9 0.0028 4 out of 8 CO 2 laser analyses

0.0024

0.0020

< 0.0016

0.0012

0.0008

0.0004

0.0000 0.0 1.0 2.0 3.0 4.0 5.0 ^®Ar/^°Ar Figure 6.01. 40A.r/39Ar isochron for laser fusion of glass from the Obsidian Cliff flow. Errors are reported at la. Arrow on ^éAr/^^^Ar axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

OC - 5 - glass 400 Total gas age = 105.6 ± 1.0 ka

40 60 80 100 % ^®Ar Released Figure 6.02. '^OAr/^^Ar age spectrum for step heating of glass from the Obsidian Cliff flow. Errors are reported at la.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. GR - 2 - sanidine

0.0032 Isochron age = 118.0 ± 10.0 ka 40/36 = 336.0 ± 14.0 0.0028 MSWD = 1.9 4 out of 12 single crystals 0.0024

0.0020

.< 0.0016 M(D 0.0012

0.0008

0.0004

0.0000 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 ^®Ar/^°Ar Figure 6.03. 40Ar/39Ar isochron for laser fusion of sanidine from the Gibbon River flow. Errors are reported at Ict. Arrow on 36Ar/"*®Ar axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

GR - 2 - sanidine

200 Total gas age = 133.6 ± 1.0 ka

2 0) O) <

1 0 0

% 39 Ar, Released Figure 6.04. ^^^^Ar/^^Ar age spectrum for step heating of sanidine from the Gibbon River flow. Errors are reported at la.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CC - 7 - glass

0.0036 Isochron age = 391.0 ± 14.0 ka 0.0032 40/36 = 297.0 ±5.5 MSWD = 2.0 0.0028 6 out of 11 CO 2 analyses

0.0024

0.0020

0.0016

0.0012

0.0008

0.0004

0.0000 0.0 0.2 0.4 0.6 0.8 1.0 39A r/°A r

Figure 6.05. isochron for laser fusion of glass from the Cougar Creek dome. Errors are reported at Ict. Arrow on ^^Ar/^^Ar axis indicates composition of atmospheric argon. Error ellipses are shown at 2 ct.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CC - 7 - glass

650 Plateau age = 411.4 ± 1.9 ka 600 65.7% of the 39Ar released Steps 3 through 9 550

500

g 450 Plateau 0) O) 400 < 350

300

250

2 0 0 2 0 40 60 80 1 0 0 %^®Ar Released Figure 6.06. age spectrum and plateau from step heating of glass from the Cougar Creek dome. Errors are reported at 1 a.

CC - 7 - glass 0.0036 Isochron age = 378.0 ± 2.0 ka 0.0032 40/36 = 561 ±120 MSWD = 1.9 0.0028 Steps 7 through 12

0.0024

0.0020

< 0.0016

0.0012

0.0008

0.0004

0.0000 0.0 0.2 0.4 0.6 0 .8 1 .0 1 .2 '=Ar/^Ar Figure 6.07. isochron from furnace step heating of glass from the Cougar Creek dome. Errors are reported at la. Arrow on axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CC-7 - sanidine

850 Total gas age = 368.2 ± 2.0 ka 750 to 1867 ±23 ka

650

S 550

0) 450 O) < 350

250

150

0 2 0 40 60 80 1 0 0 % Released Figure 6.08. '^^Ar/^^Ar age spectrum for furnace step heating of sanidine from the Cougar Creek dome. Errors are reported at la.

CC-7 - sanidine 0.0036 Isochron age = 286.0 ± 28.0 ka 0.0032 40/36 = 373 ± 77 MSWD = 2.0 0.0028 Steps 5 through 8

0.0024

0.0020

0.0016

0.0012

0.0008

0.0004

0.0000 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

Figure 6.09. isochron for furnace step heating of sanidine from the Cougar Creek dome. Errors are reported at la. Arrow on ^^Ar/^OAr axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RF - 2 - sanidine

0.0032 Isochron age = 525.8 ± 3.3 ka 40/36 = 306.3 ± 4.5 0.0028 MSWD = 1.6 7 out of 14 single crystals 0.0024 k. < 0.0020

0.0012

0.0008

0.0004

0.0000 0.00 0.20 0.40 0.60 0.80 W * A r

Figure 6.10. isochron for laser fusion of sanidine from the Riverside Flow. Errors are reported at 1er. Arrow on axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. GH - 2 - sanidine

0.0036 Isochron age = 134.3 ± 2.6 ka 0.0032 40/36 = 295.3 ±1.5 MSWD = 1.6 0.0028 8 out of 13 single crystals

0.0024

0.0020

0.0016

0.0012

0.0008

0.0004

0.0000 0 .0 0.5 1 .0 1.5 2.0 2.5 3.0 3.5 ^®Ar/Âr Figure 6.11. ^0Ar/39j\r isochron for laser fusion of sanidine from the Gibbon Hill dome. Errors are reported at la. Arrow on ^Âr/^Âr axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

PH - 5 - sanidine

0.0032 Isochron Age = 208.1 ± 4.9 ka 40/36 = 305.0 ± 6.0 0.0028 MSWD = 2.1 8 out of 15 single crystals 0.0024

0.0020

0.0016 S 0.0012

0.0008

0.0004

0.0000 0 .00.4 0 .8 1 .2 1 .6 2 .0 ^^Ar/^Ar

Figure6 .12. ÔAr/^Âr isochron for laser fusion of sanidine from the Paintpot Hill dome. Errors are reported at la. Arrow on ^Âr/ÔAr axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LD - 1 - sanidine

0.0036 Isochron age = 226.0 ± 5.5 ka 0.0032- 40/36 = 288.0 ± 9.5 MSWD = 1.8 0.0028- 11 out of 15 single crystals

0.0024 - O 0 . 0020 - < CO 0.0016-

0 . 0012 -

0.0008-

0.0004-

0.0000 0.4 0 .8 1 .2 1 .6 2 .00.0 W ° A r Figure 6.13. isochron for laser fusion of sanidine from the Landmark dome. Errors are reported at la. Arrow on ^^Ar/'^^^Ar axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

GLM - 2 - sanidine 0.0036 Isochron age = 263.3 ± 3.4 ka 0.0032 40/36 = 299.0 ± 2.5 MSWD = 1.6 0.0028 5 out of 8 CO 2 laser analyses

0.0024

0.0020

< 0.0016

0.0012

0.0008

0.0004

0.0000 0 .00.4 0 .8 1.2 1 .6 W “Ar Figure 6.14. 4ûAr/39Ar isochron for laser fusion of multiple sanidine crystals from the Grizzly Lake mingled lava. Errors are reported at la. Arrow on ^^Ar/'^OAr axis indicates composition of atmospheric argon. Error ellipses are shown at 2 a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. GRM - 1 - sanidine 0.0036 Isochron age = 300.5 ± 2.8 ka 0.0032 40/36 = 295.9 ± 0.6 MSWD = 1.9 0.0028 8 out of 12 CO2 laser analyses

0.0024

o< 0.0020

< 0.0016 % 0.0012

0.0008

0.0004

0.0000 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 ^^Ar^Ar Figure 6.15. isochron for laser fusion of multiple sanidine crystals from the Gardner River mingled lava. Errors are reported at Ict. Arrow on axis indicates composition of atmospheric argon. Error ellipses are shown at 2 ct.

AS-1 - sanidine 0.0036 Isochron Age = 316.0 ± 2.2 ka 0.0032 40/36 = 281.0 ±4.8 MSWD = 1.4 0.0028 9 out of 14 single crystals 0.0024

^ 0.0020

0.0008

0.0004

0.0000 0 .0 0 .2 0.4 0 .6 0 .8 1 .0 1.2 1.4 ^ÂrÂr Figure 6.16. 40Ar/39Ar isochron for laser fusion of sanidine from the Appolinaris Spring dome. Errors are reported at 1er. Arrow on ^Âr/^Âr axis indicates composition of atmospheric argon. Error ellipses are shown at 2cr.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. WP - 1 - sanidine 0.0036 Isochron age = 325.8 ± 2.2 ka 0.0032 40/36 = 301.1 ± 1.7 MSWD = 1.4 0.0028 8 out of 13 single crystals

0.0024

o< 0.0020

< 0.0016

0.0012

0.0008

0.0004

0.0000 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 ''A r/^ A r Figure 6.17. isochron for laser fusion of sanidine from the Willow Park dome. Errors are reported at la. Arrow on ^^Ar/'^OAr axis indicates composition of atmospheric argon. Error ellipses are shown at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 7

GEOCHEMISTRY

Major and trace element concentrations from analysis of representative samples are

listed in Table 1 of Appendix B. LOI values listed in Table I were calculated (Table 4,

Appendix B) based on the procedure described in Chapter 4. Table 2 in Appendix B lists

the instrumental precision for major elements and Table 3 lists the instrumental precision

for trace elements.

In the following discussion of major element, trace element, and isotopic

geochemistry results, sample AS-2a represents a mafic enclave from the Appolinaris

Spring mingled lava that was analyzed for major and trace element concentrations after

observing the odd trace element signature of mafic enclave AS-1 a (see below). Enclave

AS-2a was not analyzed for its isotope geochemistry. Both enclaves were removed from

the same rhyolite host sample, AS-1.

Major Element Chemistry

Classification

Using the classification scheme of LeBas et al. (1986), all of the Obsidian Creek and

Roaring Mountain member lavas are high-silica rhyolite (> 75 wt. % SiOz) with the

exception of mafic enclaves from the Appolinaris Spring, Gardner River, and Grizzly

Lake mingled lavas and the silicic host of the Grizzly Lake mingled lava (Figure 7.01).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mafic compositions of the mingled lavas range from basaltic-andesite (AS-la, AS-2a,

and GRM-la) to andésite (GLM-2a). The silicic Grizzly Lake host (GLM-2) is dacite.

Major Elements Versus SiO?

Major element oxide abundances of the Obsidian Creek and Roaring Mountain

rhyolites show no systematic variations with increasing SiOz on Marker variation

diagrams (Figures 7.02 and 7.03). This is largely because SiOz values cover a narrow

range from 75.0 to 77.3 wt. %. There are also no systematic major element variations

with age (not shown) since major element concentrations are similar among these high-

silica rhyolites (Table 1, Appendix B).

The only samples that define significant trends with increasing SiOz (and age) are

components of mingled lavas, which range jfrom 53.0 to 65.9 wt. % SiOz. These basaltic-

andesite enclaves (AS-la and AS-2a at 316 ka and GRM-la at 301 ka) and dacite host

(GLM-2 at 263 ka) decrease in Al, Fe, Ca, Mg, Ti and P and increase in K and Na oxides

with increasing SiOz (Figure 7.02 and 7.03). The andésite component (GLM-2a) of the

Grizzly Lake mingled lava plots on the Fe, Ca, and K trends, but plots off trends for all

other major elements. It should be noted that the measured values of several minor

element oxides (TiOz, MnO, and PzOs) are so low that they may be at or near the

detection limits.

Trace Element Chemistry

In contrast to the major elements, trace elements show well-defined variations among

the Obsidian Creek and Roaring Mountain member rhyolites. Nb is cbosen as an index of

differentiation since it is a highly incompatible and immobile element that is accurately

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. measured, and shows a two-fold variation in concentration. Arrowed trends in all plots

indicate increasing or decreasing concentrations with time. It should be noted that the

Cougar Creek (CC-7) and Riverside (RF-2) lavas plot off all trends, as well as mafic

components (AS-la and GRM-la) from the Appolinaris Spring and Gardner River

mingled lavas and both mafic and silicic components (GLM-2a and GLM-2 respectively)

of the Grizzly Lake mingled lava.

Trace Elements Versus Age

In plots of trace elements versus age, Rb, Cs, Hf, Th, Nb, Ta, U, Pb, Yb, and Lu show

increases in abundance in rhyolites from 326 - 134 ka and fi'om 118 - 80 ka. Abundanees

generally show a two to three-fold increase over the 326 - 134 ka interval, and a one-half

to two-fold increase over the 118 - 80 ka interval. Plots of the large ion lithophile (LIL)

element Rb and high field strength element (HFSE) Ta versus age in Figure 7.04 are

representative of these well-defined variations.

Trace elements Ba, Sr, Zr, La, Ce, and Eu decrease in abundance over the 326 - 134

ka and 118 - 80 ka intervals. Nd concentration decreases over the 326 - 134 ka interval,

but has a relatively flat pattern over the 118 - 80 ka interval. Figure 7.04 is a plot of the

light rare earth element (LREE) La versus age.

All plots in Figure 7.04, regardless of increasing or decreasing traee element

abundances, define two trends. The first trend is defined by the Obsidian Creek member

rhyolites (filled diamonds) over the 326 - 134 ka interval. The second trend is defined the

Obsidian Creek member rhyolites (open diamonds), which erupted over the 118 - 80 ka

interval. After eruption of the youngest, most evolved Obsidian Creek member rhyolite at

134 ka, all plots show a consistent break in trend marked by a signifieant change in the

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. trace element abundance of the first erupted Roaring Mountain member rhyolite at 118

ka, which is distinctly less evolved.

Trace Elements Versus Nb

Plots of trace elements versus an index of differentiation (Nb) show trends similar to

those observed in trace element versus age plots. The incompatible element Th and the

compatible element Eu were plotted against Nb (Figure 7.05) to show examples of trends

using the same trace elements reported in the previous section. On the plot of Th versus

Nb, both the Obsidian Creek and Roaring Mountain member rhyolites show a positive

correlation with increasing Nb. On the plot of Eu versus Nb two separate potential

differentiation trends are defined, but correlate negatively with increasing Nb. Mingled

lavas and the Cougar Creek and Riverside rhyolites plot off trends as in trace element

versus age plots.

REE/Chondrite Diagrams

Figure 7.06 shows samples fi'om the Obsidian Creek member which includes two

rhyolite host lavas (AS-1 and GRM-1) and one dacite host (GLM-2) for enclaves fiom

the Appolinaris Spring, Gardner River, and Grizzly Lake mingled lavas, respectively.

The LREE’s are enriched ~ 30 - 300 times chondrite whereas the HREE’s are enriched ~

25 - 70 times chondrite. The older WP-1, AS-1, and GRM-1 samples are the least

evolved and overlap in LREE’s. Successively younger samples GLM-2, LD-1, PH-5, and

GH-2 show decreasing LREE abundance and an increasingly negative Eu anomaly. The

older and less evolved WP-1, AS-1, GRM-1, and GLM-2 samples are very similar with

steeper LREE slopes compared to the flatter LREE patterns of the younger LD-1, PH-5,

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and GH-2 samples. The HREE patterns are relatively flat and show no clear variations

with decreasing age.

Figure 7.07 displays REE/Chondrite diagrams for the Roaring Mountain member

rhyolites as originally defined (top diagram), and excludes the geochemically and

isotopically distinct (this chapter) Cougar Creek (CC-7) and Riverside (RF-2) rhyolites

(bottom diagram). The Roaring Mountain member rhyolites (GR-2, OC-5, and CS-1 as

defined in this study) are nearly identical in LREE abundanees, but show a large negative

Eu anomaly with decreasing age. HREE patterns are relatively flat and increase with

decreasing age. The Cougar Creek dome and Riverside flow (Figure 7.07, top) overlap in

LREE’s and overlap, or are enriched in HREE’s compared to the Gibbon River (GR-2),

Obsidian Cliff (OC-5), and Crystal Spring (CS-1) flows.

When plotted together (not shown), all Roaring Mountain member rhyolites are

indistinguishable from the least evolved Obsidian Creek member rhyolites (WP-1 and

AS-1 in Figure 7.06) in LREE’s. The least evolved Roaring Mountain member rhyolite

(sample GR-2) is also indistinguishable from the least evolved Obsidian Creek member

rhyolites (WP-1 and AS-1 in Figure 7.06) in HREE’s whereas successively erupted and

evolved Roaring Mountain rhyolites (samples OC-5 and CS-1) overlap with a more

evolved Obsidian Creek rhyolite (PH-5 in Figure 7.06).

In Figure 7.08 (top diagram) two basaltic-andesite enclaves (AS-la and AS-2a) are

plotted with their rhyolite host from the Appolinaris Spring mingled lava. Basaltic-

andesite enclave AS-la shows a large Eu anomaly and is enriched in LREE’s and

HREE’s relative to its rhyolite host AS-1 and to all other Obsidian Creek and Roaring

Mountain member rhyolites. A second enclave AS-2a is depleted in LREE’s, has a less

6 6

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. negative Eu anomaly, and is enriched in HREE’s relative to its host AS-1. These data

suggest that the mafie enclaves from the Appolinaris Spring mingled lava are

heterogeneous.

The basaltic-andesite enclave GRM-la from the Gardner River mingled lava shows a

depleted LREE trend, a small negative Eu anomaly, and a depleted HREE trend relative

to its rhyolite host GRM-1 (Figure 7.08). The andésite enclave GLM-2a from the Grizzly

Lake mingled lava is depleted in LREE’s and HREE’s relative to its dacite host GLM-2,

and shows no negative Eu anomaly. The andésite from the Grizzly Lake mingled lava is

enriched in LREE’s (La, Ce, and Pr) and depleted in all HREE’s relative to the basaltic-

andesite of the Gardner River mingled lava, and both are depleted in all REE’s relative to

the basaltic-andesites from the Appolinaris Spring mingled lava.

Nd, Sr, and Pb Isotope Data

Nd-Sr Isotope Data

A plot of SNd versus ^^Sr/^Sr; shows that ^^Sr/^Sr, for the Obsidian Creek and

Roaring Mountain member rhyolites are significantly heterogeneous, ranging 0.70896 to

0.71461 (Figure 7.09). The basaltic-andesite and dacite samples from the mingled lavas

all plot at -0.7060, whereas andésite from the Grizzly Lake mingled lava plots at

-0.7065.

Epsilon Nd values are tightly constrained and range between - -13.1 to -12.1

(*'*^Nd/*'*'*Nd = 0.51196 to 0.51202) among the Obsidian Creek and Roaring Mountain

member rhyolites with the exception of the Cougar Creek dome (CC-7, SNd = -15.8)

(Table 7.01 and Figure 7.09). Basaltic-andesites (GRM-la and AS-la) and dacite (GLM-

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2) from the Gardner River, Appolinaris Spring, and Grizzly Lake mingled lavas

respectively, exhibit SNd values of ~ -1 0 . 8 to -8 .0 , which are slightly more radiogenie than

the rhyolites. Andésite (GLM-2a) from the Grizzly Lake mingled lava is much less

radiogenic at SNd =-23.8.

In Figure 7.10, an oscillatory pattern in (*’Sr/^^Sr)i with decreasing age is observed

where the Obsidian Creek member rhyolites show a decrease in (®^Sr/*^Sr)j from ~ 0.713

- 0.710 over the 326 - 301 ka interval, which is followed by an increase to -0.713

approximately 75 ka later. From this point there is a decrease to - 0.711 approximately

74 ka later, which is followed by an increase from - 0.711 - 0.715 over the 134 - 106 ka

interval. This is followed by a final decrease to the lowest rhyolite value of - 0.709

approximately 26 ka later.

6 8

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Pb Isotopic Data

Plots of ^‘^^Pb/^*^‘*Pb and ^‘^^Pb/^°'’Pb versus ^°^Pb/^^'*Pb show a main cluster (solid

oval) that includes Obsidian Creek and Roaring Mountain member rhyolites and basaltic-

andesite and dacite from the Gardner River and Grizzly Lake mingled lavas respectively

(Figure 7.11). For this main cluster, ^^^Pb/^^'^Pb values range 17.143 to 17.489,

207pb/204pb values range 15.533 to 15.606, and ^°^Pb/°^Pb values range 38.064 to 38.463.

The Cougar Creek (CC-7) and Riverside (RF-2) lavas from the Roaring Mountain

member plot well outside (dotted oval) the main cluster due to lower ^®^Pb/^*^'*Pb (16.591

to 16.658) and ^"^Pb/^'*Pb (15.440 to 15.442) values (Figure 7.11). For these units,

208pb/204pb yalues range from 38.396 to 38.463, which are similar to the upper limit of

the main cluster values.

Separate from the main cluster and Cougar Creek and Riverside rhyolites are two

samples with Pb isotopic compositions that are either more or less radiogenic. Sample

AS-la, a basaltic-andesite enclave from the Appolinaris Spring mingled lava, is more

radiogenic than the main cluster in ^°^b /^P b (17.716), ^°W^°^Pb (15.786), and

208pb/204pb ( 3 8 5 1 7 ) values. Sample GLM-2a, andésite from the Grizzly Lake mingled

lava, is less radiogenic than the other lavas in ^^^Pb/^^'^Pb (16.060), ^^^Pb/^'^Pb (15.273),

and ^°W^°^Pb (36.516).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■DCD O Q. C g Q.

■D CD

C/) 3o' O

Table 7.01. 8 206pb/204pb 207pb/204pb 208pb/204pb Sample ^Sr/^^Srm ^^Sr/^^Sn "hsid/'^^d %d (O' GH-2 0.71249 ± 2 (171114 ±11 0.51198 ± 1 -12.9 17.163 ± 17 15.533 ±23 38.153 ±76 PH-5 0.71458 ± 2 0.71283 ± 15 0.51196± 1 -13.1 17.265 ± 17 15.557 ±23 38.121 ±76 LD-1 0.71444 ± 2 0.71278 ±14 0.51201 ± 1 -12.3 17.273 ± 17 15.579 ±23 38.191 ±76 GLM-2 0.70603 ± 2 0.70603 ± 2 0.51209 ± 1 -10.7 17.399 ± 17 15.605 ±23 38.203 ± 76 3. 3" GLM-2a 0.70648 ± 2 0.70648 ± 2 0.51142 ± 1 -23.7 16.060 ± 16 15.273 ±23 36.516 ±73 CD GRM-1 0.71035 ± 2 0.71028 ± 2 0.51200 ± 1 -12.4 17.466 ± 17 15.591 ±23 38.143 ±76 ■DCD O GRM-la 0.70587 ± 2 0.70586 ± 2 0.51223 ±1 -8 . 0 17.424 ± 17 15.573 ±23 38.064 ±76 Q. C AS-1 0.71218 ± 2 0.71212 ±2 0.51202 ± 1 - 1 2 .1 17.486 ± 17 15.581 ±23 38.125 ±76 a o AS-la 0.70590 ± 2 0.70590 ± 2 0.51208 ± 1 - 1 0 . 8 17.716± 18 15.786 ±24 38.617 ±77 3 ■D o WP-1 0.71322 ± 2 0.71313±2 0.51201 ± 1 -12.3 17.489 ± 17 15.591 ±23 38.125 ±76 O

CD CS-1 0.70906 ± 2 0.70896 ± 2 0.51198± 1 - 1 2 . 8 17.143 ± 17 15.536 ±23 38.259 ±77 Q. OC-5 0.71503 ± 2 0.71461 ± 2 0.51198 ± 1 - 1 2 . 8 17.208 ± 17 15.601 ±23 38.463 ± 77 GR-2 (171319 ± 2 0.71319± 3 0.51200 ± 1 -12.5 17.143 ± 17 15.553 ±23 38.348 ±77 ■D CC-7 0.71204 ± 2 0.71185 ± 2 0.51182 ± 1 -15.9 16.591 ±17 15.440 ±23 38.463 ± 77 CD RF-2 0.71269 ± 2 0.71087 ± 12 0.51197± 1 -13.0 16.658 ±17 15.442 ±23 38.396 ±77

C/) 2 C/) All errors reported at cr and shown in the last decimal place(s). Pb errors are based on KU in-house common Pb standard NBS 981, Sr errors are based on standard NBS 987, and Nd values are adusted using a KU in-house standard that is referenced to La Jolla Nd. 87 Sr/ 86 Sr, values calculated using measured 87 Sr/ 86 Sr values, 87 Rb/ 86 Sr values, and 40 All . ,39 Ar . ages. 16

14 Phonolite

Tephri- 1 2 phonolite Trachyte

o Phono- Tephrite TrachydaciteTrachy- andesite Rhyolite 8 Basaltic GLM-2 oCsl trachy- (0 andesite Trachy- GLM-2a 6 basalt AS-1a^ lGRM-1a Dacite 4 AS-2a Basait Andésite 2 Basaltic andésite

0 50 5545 60 65 70 75

S i 0 2

Figure 7.01. Total alkalis (Na2 Û + K2O) versus Si0 2 classification diagram for the Obsidian Creek (♦) and Roaring Mountain (^) member extracaldera rhyolites and mingled components, AS-la (A), AS-2a (A), GRM-la (A), GLM-2 (•), and GLM-2a (A) of the Appolinaris Spring, Gardner River, and Grizzly Lake mingled lavas respectively (classification after LeBas et al., 1986). All analyses normalized to 100% anhydrous and expressed in wt. %.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18

16 oCO CM 14

■* 7 o 0) O) 5

1 60 80

Figure 7.02. Harker variation diagrams plotting AI2 O3 , FeO* (total Fe as FeO), CaO, and MgO versus Si0 2 . Ail major elements are normalized to 100% anhydrous and expressed as wt. %. Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6

O 4 CM 3

4 0.2 O 3 ▲ ▲ m S' z 2 0.1 1

0 0.0 50 60 70 80 50 80 Si0 2 “ SiOa ™

Figure 7.03. Harker variation diagrams plotting K2 O, Na2 Û, Ti0 2 , and P 2 O5 versus Si0 2 . AU major elements are normalized to 100% anhydrous and expressed as wt. %. Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 500

400

E 300 Q. RF-2 CC-7 5 200

* GLM-2 1 0 0 GRM-1 a A 4 à A S -1 a GLM-2a ,

5 RF-2 Cl 4

2 ^GLM-2 vcC-7 AS-13 1 GRM-la A G L M -2a 0 CC-7 RF-2 70

g . 50 GLM-2 a.

CD 40 2\GLM-2a

30 ^ G R M -la

10 0 1 0 0 200 300 400 500 600 Age (ka)

Figure 7.04. Age versus trace elements (Rb, Ta, and La) plots showing increasing/ decreasing trends with age for the Obsidian Creek and Roaring Mountain member rhyolites. Mingled lavas and the Cougar Creek (CC-7) and Riverside (RF-2) rhyolites from the Roaring Mountain member commonly plot off trends. Arrows indicate general trends, not models. Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40

CC-7 30

Q. CL RF-2 20

# GLM-2

10 ^ AS-1 a GRM-la

AS-13

1.5 GRM-la

E GLM-2a Q. Q. GLM-2 3 1.0 LU

RF-2 0.5 CC-7

0 0 20 40 60 80 Nb (ppm)

Figure 7.05. Trace elements Th and Eu versus Nb. Arrows indicate trends for the Obsidian Creek and Roaring Mountain member rhyolites. Mingled lavas and the Cougar Creek (CC-7) and Riverside (RF-2) rhyolites from the Roaring Mountain member plot off trends. Arrows indicate general trends, not models. Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100

0 ) "O c o Ü ^ 10 E CO C/5 q}] WP-1

□ AS-1

# GRM-1

A GLM-2

0 LD-1 ★ ph-5 Young > i ^ GH-2

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 7.06. REE/Chondrite plots for the Obsidian Creek member rhyolites, including dacite sample GLM-2 from the Grizzly Lake mingled lava. Direction of arrow indicates decreasing age. Chondrite values from Sun and McDonough (1989).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 I I r 1— r 1— I— I— I— I— I— r

100 a ■D C o x z

0 ) Q. 10 O CC-7 E CD (/) OC-5 Young

J L I I I I I I I I j L La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

1 0 0

Old I ^ GR-2 Q. I O OC-5

Young CS-1

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 7.07. REE/Chondrite plots (Sun and McDonough, 1989) for the Roaring Mountain member rhyolites. Top diagram shows all five rhyolite domes/flows as originally assigned to the Roaring Mountain member. Bottom diagram eliminates the older Riverside flow (RF-2) and Cougar Creek dome (CC-7), which are geochemically and isotopically distinct, and displays only the younger eruptive units. Arrows indicate decreasing age.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. # 100 ■D

“ ^ AS-1 (rhyolite) ▼

A A S -la (basaltic-andesite enclave)

- AS-2a (basaltic-andesite enclave)

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

1 — r 1 — r

(D 100 X 3 C 0 1 Q. E 03 co ^ GRM-1 (rhyolite) A GRM-la (basaltic-andesite enclave) 10 r GLM-2 (dacite) I È k GLM-2a (andésite enclave) I I I I I I I I La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 7.08. REE/Chondrite plots (Sun and McDonough, 1989) for the Appolinaris Spring, (top), Gardner River (bottom), and Grizzly Lake (bottom) mingled lavas.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 T 1------1------1------1------1------1------T

Bulk Earth

j^RM -1 a

GNd - 1 0 I GLM-2 AS- la

-20

A GLM-2a

I I I I J I I L -30 0.705 0.707 0.709 0.711 0.713 0.715 (87Sr/86Sr)i

Figure 7.09. Plot of whole rock (87Sr/86Sr)i versus epsilon Nd. The Obsidian Creek and Roaring Mountain member rhyolites have s^d o f-12 and variable (87Sr/86Sr)j values. The Riverside flow overlaps at -13 and the chemically distinct Cougar Creek dome has a lower value of -15. Basaltic-andesites (GRM-la and AS-la) from the Gardner River and Appolinaris Spring mingled lavas, and dacite (GLM-2) from the Grizzly Lake mingled lava have of- 8 to -12, whereas andésite (GLM-2a) from the Grizzly Lake mingled lava has a distinctly lower value of -23. The (87Sr/86Sr)j values for basaltic-andesites, andésite, and dacite have lower (87Sr/86Sr)j values (~ 0.706) than the rhyolites. Symbols as in Figure 7.01. Composition of "bulk" Earth shown for reference.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.719

0.717

0.715

0.713 CC-7 RF-2 0.711

0.709

0.707 A ÀL 0.705 0 1 0 0 200 300 400 500 600 Age (ka)

Figure 7.10. (87Sr/86Sr)j versus age plot. (87Sr/86Sr)j shows no systematic enrichment with increasing age for the Obsidian Creek or Roaring Mountain member rhyolites. Basaltic-andesite, andésite, and dacite from the mingled lavas plot at ~ 0.706 in (87Sr/86Sr)j Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15.8 1 1 ▲ AS 1-3 15.7 -

15.6 -

é P - 207pb/204pb 15.5

15.4 —

15.3 A GLM-2a 1 I 1 15.2 16 17 18 206pb/204pb

39 AS-13

20Spb/204pb

36 16 17 18 206pb/204pb

Figure 7.11. Plots of 207pb/204pb and 208pb/204pb versus 206pb/204pb values for the Obsidian Creek and Roaring Mountain member lavas. Plots show a main cluster (solid oval) of Obsidian Creek and Roaring Mountain member rhyolites including basaltic- andesite and dacite from the Gardner River and Grizzly Lake mingled lavas respectively. The Cougar Creek and Riverside lavas from the Roaring Mountain member plot well outside (dashed oval) the main cluster in both diagrams. Basaltic- andesite from the Appolinaris Spring mingled lava shows enriched values compared to the main cluster and andésite from the Grizzly Lake mingled lava shows depleted values relative to the main cluster. Symbols as in Figure 7.01.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 8

PETROGENESIS

Petrogenetic models were done to identify the processes involved in the petrogenesis

of the extraealdera rhyolite magmas. These models were used to determine potential

crustal source rocks of the extraealdera rhyolite magmas and to identify if the observed

variation of composition with age is caused by fractional crystallization (EC), if magma

mixing generated the mingled lavas, and/or if contamination resulted from assimilation of

crustal rocks beneath the Yellowstone Plateau.

The term magma “mingling” as used in this study refers to the intimate interaction of

two compositionally distinct magmas without physical mixing or “hybridization”

occurring, which is a term reserved to describe the mixing of two magmas to produce a

new magma. The discrete enclaves formed during magma mingling can be attributed to

the quenching of hotter magma in the cooler rhyolite/dacite host lavas (Murphy et al.,

2000).

Hybrid Source Models

Figure 8.01 shows a SNd versus ^’Sr/^^Sq diagram that plots the isotopic compositions

of the rhyolites (Table 7.01), several representative Swan Lake Flat basalts (Bennett, in

progress), and crustal xenoliths from the Snake River Plain (Leeman et al., 1985) and

Crazy Mountains (Dudas et al., 1987). The rhyolites plot at intermediate isotopic

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. compositions between the basalts and xenoliths and can therefore be modeled as hybrids

between such end members.

Feasible crustal end members would be expected to have ^^Sr/^Sq between -0.710

and 0.730 and Enci values between — 25 and - 45. Several xenoliths from the Snake

River Plain and one from the Crazy Mountains have compositions within this range

(Chapter 9) and were therefore selected for mixing models presented below. All other

xenoliths from the Crazy Mountains (not shown) have ^^Sr/^^Sr, values (0.7053 - 0.7086)

that are less radiogenic and End values (-9.5 and -19.3) that overlap with the rhyolites.

These intermediate isotopie compositions are not considered since they may represent

crustal rocks that are already hybridized. Snake River Plain xenoliths less radiogenic

(87sr/86sri < 0.7053) than Swan Lake Flat basalts were also ruled out as end- members

because such compositions could not account for the elevated ^^Sr/^Sq compositions of

the rhyolites.

The Quaternary age of the Yellowstone basalts suggests that they were the most

likely magmas to have been injected into the deep crust resulting in fractional

crystallization, partial melting of crustal rocks, and generation of potential hybrid

reservoirs. The Swan Lake Flat basalts (350 - 209 ka, Bennett, in progress), which

precede and overlap the eruption ages of the rhyolites (Chapter 9), may represent the

most feasible mantle end member.

Five crustal end-members were modeled with Swan Lake Flat basalt to determine the

relative crust-basalt contributions needed to generate a hybrid parental material similar to

the extraealdera rhyolites. The crustal end members consist of four Snake River Plain

xenoliths and one xenolith from the Crazy Mountains. Of the four representative Swan

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Lake Flat compositions used in this study, sample SLF-1 was chosen as the basalt end-

member in all models since other samples show evidence from Pb isotopes for crustal

assimilation (Chapter 8 ). Table 8.01 gives the measured Nd and Sr isotopic values of the

end member compositions and lists the weight fractions of crust and basalt for each

model.

Model results are shown on Figure 8.01 as five mixing arrays that intersect the

isotopic compositions of the least-evolved extraealdera rhyolites. The crustal end

members used have ^^Sr/^Sq compositions that range from -0.709 - 0.725. Sample CS-1

from the Crystal Spring flow has a ^^Sr/^Sq = 0.709 indicating it is the least evolved and

thus most likely to represent the extraealdera rhyolites before they were modified in the

upper crust. However, it is possible that all rhyolites, including CS-1, have elevated

^^Sr/^Sq due to a small amount of upper crustal assimilation. The low Sr concentrations

(1-32 ppm) of the rhyolites make them extremely sensitive to minute amounts («1% )

of wall rock interaction. The upper crustal assimilation trend line in Figure 9.01 shows

the potential effects upper crustal assimilation, however, it has not been modeled since

the extraealdera rhyolites show no systematic enrichment in ^’Sr/^^Sq with time (Figure

7.10).

Using the Swan Lake Flat basalt (SLF-1) and the selected crustal end-members listed

in Table 8.01, mixing models 2, 3, and 5 predict crustal contributions of -30 - 60% and

are preferred over other models since they most closely intersect the isotopic composition

of CS-1. Model 4 yields an acceptable result that predicts a greater crustal contribution

of -70%, but intersects the intermediate isotopic composition of evolved rhyolites. Model

1 results in a ^^Sr/^Sr, value lower than values observed for the rhyolites (0.709 - 0.714)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and predicts a crustal contribution of only -10% due to the high Nd concentration (85

ppm) of the crustal end member.

Rayleigh Fractionation Models

Fractional crystallization dominates the chemical evolution of the Obsidian Creek and

Roaring Mountain member rhyolites as evidenced by the various trace element fractional

crystallization models presented below. Models were calculated using the Rayleigh

fractionation equation, Cl/Cq = where Co = parent liquid. Cl = daughter liquid, F =

weight fraction of melt remaining, and D = bulk distribution coefficient. The

fractionation models assume that crystallizing phenocrysts were separated from the melt

in the modal proportions determined from pétrographie analysis (Chapter 5). Because

abundances of accessory phases (zircon, allanite, magnetite, and apatite) were not

constrained hy point counting, hypothetical abundances were used and adjusted as

needed.

Since the Rayleigh equation depends on the bulk distribution eoefficient (D), it is

important that appropriate Kd values are used in fractionation modeling. The list of Kd’s

presented in Table 8.02 are averages derived from a eompilation of studies (Mahood and

Hildreth, 1982; Stix and Gorton, 1990; Ewart and Griffin, 1994; and Streck and Grander,

1997) on trace element partitioning for several high-silica rhyolite centers.

Fractionation models that use incompatible trace elements such as Ta or Th as an

index of fractionation produce results similar to the Nb models. Analytical uncertainties

(la) for analyzed trace elements used in fractionation models range from 0.9 - 9.50 %

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and average ~ 2.4 %; therefore, daughter magmas that plot within 2a of calculated

fractionation trends are considered a good fit to the model.

The low Sr concentrations (Chapter 7) in the rhyolites and very high ^^Sr/^Sr ratios

of Archean crust (Chapter 9) make the rhyolites very sensitive to any assimilation.

However, the possibility that country rock was assimilated during fi-aetional

crystallization and contributed to differentiation is considered unlikely since (^^Sr/^Sr);

does not systematically increase with decreasing age as shown in Figure 7.10.

Fractional Crystallization Model for the

Obsidian Creek Member Rhvolites

A fractional crystallization model calculated for the Obsidian Creek member rhyolites

using the Willow Park composition (WP-1) as the parental magma show a good fit

between the calculated daughter (C l) compositions and the actual evolved daughter

compositions AS-1, LD-1, PH-5, and GH-2 for most trace elements considered. A best-fit

model was produced using the modal abundances (Table 5.01) of quartz (12.8%),

sanidine (82.6%), plagioclase (2.7%), and clinopyroxene (1.8%) for the fractionating

parent magma WP-1. Accessory phase abundances used in the model include those for

zircon (0.0 - 0.5%), allanite (0.01 - 0.8%), magnetite (0.5 - 3.0 %), and apatite (0.1 - 0.5

%). Figures 8.02 and 8.03 show models that plot Nb against compatible LREE La and

MREE Eu, and incompatible HFSE’s Ta and Th. Trace elements that do not show a good

fit when plotted against Nb include Y, Sr, Ba, Rb, Sm, Yb, Hf, and Lu. Poor fit of data to

the model may be a result of using partition coefficients that are not appropriate for these

magmas as discussed previously. Also, mobile LIL’s such as Sr, Ba, and Rb may not be

present in their original abundances causing poor fit of these data to the model.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. All diagrams presented in Figures 8.02 and 8.03 plot the parent magma (WP-1, 326

ka) with successively erupted Appolinaris Spring (AS-1, 316 ka), Gardner River (GRM-

1,301 ka). Grizzly Lake (GLM-2, 263 ka). Landmark (LD-1, 226 ka), Paintpot Hill (PH-

5, 208 ka), and Gibbon Hill (GH-5, 134 ka) domes/flows. The rhyolite (GRM-1) and

dacite (GLM-2) hosts of mingled lavas plot slightly off the fractionation model

suggesting that other processes may be involved in their petrogenesis. Mixing between

basaltic and rhyolitic magma, as presented in the following sections, can explain the

GLM-2 composition, but cannot explain the less evolved GRM-1 composition. Figures

8.02 and 8.03 show that rhyolite host lava AS-1 overlaps with parent WP-1 at 2a and

may represent either a nearly identical magma composition, or a product of slight

fractionation of WP-1.

Figure 8.02 shows the fractional crystallization model on Nb versus La and Eu plots.

La is strongly partitioned into the accessory phase allanite, and moderately partitioned

into accessory phases zircon, magnetite, and apatite (Table 8.02). Although Eu is strongly

partitioned into allanite and moderately partitioned into sanidine, plagioclase, zircon,

magnetite, and apatite, it is the high abundance of fractionating feldspars that controls the

Eu fractionation trend. Daughter magmas AS-1, LD-1, PH-5, and GH-2 plot remarkably

well on the calculated fractionation model suggesting that simple fractional

crystallization can explain the geochemical variations observed in these rhyolites. GRM-

1 does not show good fit to the model on Nb vs. LREE plots (eg: Nb vs. La), however it

does fit to the model on the Nb vs. Eu plot, which suggests it may have chemieally

evolved by feldspar fraetionation. The fraetional crystallization model indicates that

daughter compositions AS-1, LD-1, PH-5, and GH-2 can be produced by ~ < 5%, -40 -

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42%, -45%, and -65 - 75% fractional crystallization of parent WP-1, respectively.

GRM-1 can be produced by - 8 % fractionation of WP-1 based on the Nb vs. Eu plot only.

Thus, overall it does not fit the model.

Figure 8.03 shows the fractionation model on Nb versus Th and Ta plots. Both Th

and Ta are incompatible in the bulk fraetionating phenocryst assemblage. AS-1 is

indistinguishable from WP-1 whereas daughters LD-1, PH-5, and GH-2 can be produced

from -40%, -45 - 50%, and -60 - 75% fraetional crystallization of parent WP-1,

respectively. Again, GRM-1 is slightly less evolved in Nb, and nearly overlaps with AS-1

and WP-1.

Fractional Crvstallization Models for the

Roaring Mountain Member Rhvolites

A fractional crystallization model calculated for the Roaring Mountain member

rhyolites (excluding the Cougar Creek dome and Riverside flow) using the Gibbon River

eomposition (GR-2, 118 ka) as the parental magma shows a remarkable fit of the

ealculated daughter (C l) compositions to the Obsidian Cliff (OC-5, 106 ka) and Crystal

Spring (CS-1, 80 ka) rhyolites for all traee elements considered. As discussed previously,

the Cougar Creek dome and Riverside flow are spatially, temporally, and geoehemically

distinet from the other rhyolites, and are therefore not eonsidered in the following

geochemical model. A best-fit model was produced using the modal abundanees (Table

5.03) of quartz (13.9%), sanidine (78.5%), plagioelase (3.8%), clinopyroxene (1.3%), and

orthopyroxene (2.5%) from fractionating parent Gibbon River magma (GR-2).

Abundances of accessory phases used in the model inelude those for zireon (0.0 - 0.2 %),

allanite (0.0 - 0.1 %), magnetite (0.1 %), and apatite (0.0 - 0.4%).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figures 8.04, 8.05, and 8.06 show the fractional crystallization model on Nb versus

LREE (La), MREE’s (Sm and Eu), HREE (Yb), HFSE’s (Hf and Ta), and LIL’s (Rb and

Ba) plots. The compatible behavior of La indicates that allanite was a fractionating phase

of the Roaring Mountain magmas, and the compatible behavior of Eu and Ba indicate the

fractionation of feldspars. The fractional crystallization model is consistent with -45 -

50% fractionation of Gibbon River magma to produce the daughter compositions of OC-

5 and CS-1.

Mixing Models for the Mingled Lavas

Ample field and pétrographie evidence for mingling and mixing of silicic and mafic

magma promoted the examination of major element, trace element, and isotopic mixing

models. Mingling between silicic and mafic magma occurred in the Appolinaris Spring

(316 ka), Gardner River (301 ka), and Grizzly Lake (263 ka) lavas as well as in the

younger Crystal Spring lava (80 ka), however no mafic component chemistry was

obtained on Crystal Spring enclaves due to their small size. Models discussed below

suggest the Appolinaris Spring and Gardner River mingled lavas are composed of an end

member rhyolite host and hybrid basaltic-andesite enclaves produced by mixing between

rhyolite and basalt end-members. The Grizzly Lake mingled lava is composed of a hybrid

dacite host, however the analyzed andésite enclave GLM-2a cannot be modeled by

mixing between rhyolite and basalt end-members. Sample GLM-2a was thought to

represent the abundant mafic clots/enclaves within dacite host GLM-2, however

petrography (Chapter 5) has shown that it is distinct in its mineralogy. The origin of

GLM-2a will thus be discussed later.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In all models Swan Lake Flat basalt compositions (Bennett, in progress) were chosen

as the mafic mixing end members because of their close spatial and temporal association:

Specific Swan Lake Flat basalt compositions used for major element mixing models were

determined based on the isotopic mixing models as described in a following section.

Major Element Mixing Models

Acceptable models to produce the Gardner River basaltic-andesite enclaves (GRM-

la) and the Grizzly Lake dacite (GLM-2) were obtained using end members represented

by Gardner River rhyolite (GRM-1) and Swan Lake Flat basalt (SLF-2 and SLF-5).

Acceptable models for the Appolinaris Spring basaltic-andesite enclaves (AS-la) were

obtained using end members represented by Appolinaris Spring rhyolite (AS-1) and

Swan Lake Flat basalt (SLF-1 and SLF-14).

Basaltic andésite (GRM-1 a) can be produced by mixing -70% SLF-2 basalt with

-30% GRM-1 rhyolite and GLM-2 dacite can be produced by mixing -40% SLF-2 basalt

with -60% GRM-1 (Figures 8.07 and 8.08). It should be noted that identical mixing

models (not shown) were produced using the major element concentrations of basalt

sample SLF-5.

Acceptable models for the Appolinaris Spring enclaves (AS-la) are illustrated in

Figures 8.09 and 8.10. These models show that AS-la can be produced using -80% SLF-

1 basalt with -20% AS-1 rhyolite. A model using the major element concentrations of

SLF-14 produced identical results (not shown). It should be noted that a second analyzed

enclave, AS-2a, is successfully modeled using -90% Swan Lake Flat basalt (not shown).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Trace Element Mixing Models

Trace element concentrations are plotted for end members GRM-1 and SLF-2 with

enclave GRM-1 a (Figure 8.1 la) and dacite GLM-2 (Figure 8.12a), and for end members

AS-1 and SLF-1 with enclave AS-la (Figure 8.13a).

The Gardner River basaltic-andesite (GRM-1 a) is produced by mixing -75% SLF-2

basalt with -25% GRM-1 rhyolite (Figure 8.1 lb). These percentages agree with major

element models (Figure 8.07 and 8.08). A second mixing model using the trace element

concentrations of SLF-5 produced an indistinguishable result (not shown).

The Grizzly Lake dacite (GLM-2) can be produced by mixing -40% SLF-2 basalt

with -60% GRM-1 rhyolite (Figure 8.12b). These percentages agree with major element

models (Figures 8.07 and 8.08). A second mixing model using the trace element

concentrations of SLF-14 produced indistinguishable results (not shown).

Although the major element mixing models can consistently reproduce the

Appolinaris Spring enclave AS-la chemistry, trace element concentrations cannot be

successfully modeled. This is because AS-la is enriched in REE’s relative to both end

members AS-1 and SLF-1 (Figure 8.13b). The other analyzed enclave (AS-2a) also

cannot be modeled using these end members.

An acceptable model to produce AS-2a was attained by mixing SLF-1 with the

composition of a granitic xenolith (Bennett, in progress) brought to the surface in SLF-1

basalt. Major and trace element chemistry results (Bennett, in progress) show that the

xenolith is a suitable end member since its REE composition is enriched relative to the

mafic enclave (AS-2a) (Figure 8.14a). AS-2a can be modeled using 40% SLF-1 with

60% granitic xenolith (Figure 8.14b).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Isotopic Mixing Models

Major and trace element mixing models show that enclaves in the Gardner River

mingled lava and the dacite host of the Grizzly Lake mingled lava could both be derived

by mixing of the Gardner River rhyolite with the spatially and temporally associated

Swan Lake Flat basalt. Similarly, major element mixing models show that enclaves in the

Appolinaris Spring mingled lava could be produced by mixing the host rhyolite with

Swan Lake Flat basalt.

Isotopic mixing models were calculated using a variation of the DePaolo mixing

equation (1981) and input values of Sr, Nd, and Pb isotopic ratios with their respective

whole-rock element concentrations (Chapter 7).

^^Sr/^Sq versus SNd Mixing Models

The basaltic-andesite enclave GRM-1 a can be produced by mixing -70% SLF-2 and

-30% GRM-1 (Figure 8.15). The dacite host GLM-2 of the Grizzly Lake mingled lava

can similarly be modeled using -30 - 40% SLF-1, SLF-2, or SLF-14 and -60 - 70%

GRM-1. These mixing proportions are in good agreement with those from the major and

trace element mixing models.

Basaltic-andesite enclave AS-la from the Appolinaris Spring mingled lava is most

closely modeled using SLF-1 or SLF-14 and host rhyolite AS-1. The mixing models

(Figure 8.16) show that AS-la can best be modeled using -35% SLF-1 or SLF-14 and

-65% AS-1. The proportion of basalt needed to produce the isotopic composition of AS-

la is much less than that needed in the major element models (-80%).

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Pb Isotope Mixing Models

It should be noted that Swan Lake Flat basalt samples SLF-2 and SLF-5 have low Pb

isotopic ratios (Figure 8.18), which may reflect crustal contamination, whereas more

primitive samples SLF-1 and SLF-14 (Figure 8.20) are relatively enriched in their Pb

isotopic ratios. It would therefore be impossible for the basaltie-andesite enclave GRM-

la and dacite GLM-2 to be modeled using rhyolite GRM-1 with SLF-1 or SLF-14. Both

GRM-1 a and GLM-2 are best modeled by using -75 - 80% SLF-5 and -20 - 25% GRM-

1 (Figure 8.17). These percentages are in close agreement with the Sr and Nd isotopic

models. The fact that GRM-1 a and GLM-2 plot on Pb mixing lines with the basalt

samples (SLF-2 and SLF-5) affected by crustal contamination, eliminates samples SLF-1

and SLF-14 as potential end members for the previous major and trace element models.

The Pb isotopic compositions of enclave AS-la from the Appolinaris Spring mingled

lava indicate that the more primitive Swan Lake Flat basalt samples (SLF-1 and SLF-14)

would be the most suitable end members. Figures 8.19 and 8.20 however show that SLF-

1 and SLF-14 are not radiogenic enough to successfully model the isotopic composition

of AS-la. In Figure 8.19, AS-la shows a poor fit to the models because its ^°’Pb/^'^'*Pb

composition is more radiogenic than the basalt end members. Models in Figure 8.20

produce more acceptable fits than preceding models, however AS-la falls off mixing

lines by more than ± 2a. The isotopic composition of enclave AS-la would require a >

95% contribution from the Swan Lake Flat basalts (Figure 8.20), however these

percentages are not in agreement with the predicted contributions (-35%) from the

^^Sr/^Sr, versus cncI models.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ■DCD O Q. C g Q.

■D CD

C/) 3o' O

(O'

Table 8.01. Nd and Sr hybrid source models. End-member Compositions ~ Wt.% contribution of 3 Mixing end-members Rock Type Sr (ppm) Nd (ppm) *’Sr/**Sr crustal mixing end-member 3" CD Swan Lake Flat (SLF-1) basalt 356 18 0.70540 0 CD ■D Model 1. COM SI-1 norite 70 85 0.71795 -40 1 0 % O Q. Model 2. KM FL8X gneiss 760 7 0.70976 -34 60% C a Model 3. COM 70-40 opdalite 324 10 0.71525 -43 40-45% o 'O 3 U) 8 ■D Model 4. SKBll chamockite in gneiss 338 0.71362 -23 70% O Model 5. SK 73-68X norite 175 22 0.72550 -34 30%

CD Q. Swan Lake Flat basalt (SLF-1) was mixed with crustal xenoliths from Craters of the Moon (COM) and Spencer-Kilgore (SK) localities of the Snake River Plain area, and from the Crazy Mountains (KM) to produce 5 hybrid source models. The crustal contributions necessary to model the extracaldera rhyolites shown in Figure 9.01 are listed here as an approximate weight percent. ■D CD Swan Lake Flat basalt data from Bennett (in progress) and crustal xenolith data from Leeman et al. (2002) and Dudas et al. (1987). Best-fit models are in bold font. C/) C/) Element San Plag Bio Cpx Opx Zir Allan Mgt Ap Sc 0.04 0.17 15.50 131 2 2 . 0 0 68.65 55.85 8.81 0 . 0 0 Rb 0.48 0.25 4.20 0.00 0.01 0.00 0.19 0.04 0 . 0 0 Sr 4.00 11.80 7.20 0 . 0 0 0.17 0 . 0 0 1.80 0.09 2 . 0 0 Y 0.04 0.05 2.40 0 . 0 0 1 . 1 0 60.00 95.50 3.21 40.00 Zr 0 . 1 0 0.09 0.50 2.31 0.05 6400.00 0.29 0.62 0 . 1 0 Nb 0.09 0.14 9.10 0 . 0 0 031 50.00 1.70 2.50 0 . 1 0 Cs 0 . 0 1 0.05 2.30 0.00 0.00 0.00 0.00 0 . 0 0 0 . 0 0 Ba 4.70 13.05 5.37 0 . 0 0 0.90 0.00 0.00 0.00 2.00 La 0.06 0 . 1 0 3.18 17.60 14.40 16.90 2595.00 21.50 2 0 . 0 0 Ce 0 . 0 2 0.06 2.80 10.25 12.30 16.75 2279.00 17.85 35.00 Sm 0 . 0 1 0 . 0 1 1.55 10.36 7.87 14.40 867.00 8.14 63.00 Eu 4.04 4.75 0.87 7.47 2.85 16.00 1 1 1 . 0 0 4.05 30.00 Tb 0 . 0 1 0 . 0 2 1.05 5.50 5.50 37.00 273.00 4.75 2 0 . 0 0 Yb 0 . 0 0 0 . 0 2 0.54 5.53 2.35 527.00 30.75 1.34 25.00 Lu 0.03 0 . 0 2 0.61 7.08 2.70 641.00 33.00 1.40 25.00 Hf 0 . 0 2 0.04 0.60 1.51 0 . 0 0 3193.00 18.90 4.41 0 . 1 0 Ta 0 . 0 2 0.05 1.34 0.91 1.13 47.50 3.15 0.87 0 . 0 0 Th 0 . 0 0 0.03 1.23 2.34 6.53 76.00 538.60 7.36 2 . 0 0 U 0 . 0 2 0.15 0.17 0.58 0.28 0 . 0 0 14.30 0.85 0 . 0 0 Pb 0.75 0.70 2 . 1 0 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0.80 0 . 0 0

Sources: Mahood and Hildreth (1983) Stix and Gorton (1990) Ewart and Griffin (1994) Streck and Grunder (1997)

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Swan Lake Flat basalts Extracaldera basaltic-andesites 10 Extracaldera andésite Extracaldera dacite Extracaldera rhyolites S LF-1 Snake River Plain crustal xenoliths Crazy Mtns. crustal xenoliths

.C S -1 -10 Upper Crustal Assimilation ► "Nd -20

-30

Upper Crust -40

Lower Crust -50 J_ 0.700 0.705 0.710 0.715 0.720 0.725 0.730 87Sr/86Sn

Figure 8.01 Comparison of Sr and Nd isotopic compositions of the Obsidian Creek and Roaring Mountain extracaldera lavas with data for Swan Lake Flat basalts (Bennett, in progress) and crustal rocks from the Wyoming Province. Field of lower to upper crustal rock compositions for the Yellowstone region is defined by Snake River Plain xenoliths from Leeman et al. (1985) and Crazy Mountains xenoliths from Dudas et al. (1987). Mixing lines represent five hybrid source models using basalt end-member SLF-1 and a variety of crustal xenolith end-members. Tick marks = 10% increments.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 a error e 60 ■

GLM-2

40 - E Q. a . 40 % \ LD-1 30 - PH-5

20 -

70%

35 45 55 6575 85

2 G error

♦ A S -1 GLM-2 WP-1

E GRM-1 Q. D. 3 lU

30% LD-1 PH-5 60% G H-2 0.0 35 556545 75 85 Nb (ppm)

Figure 8.02. Plots of Nb versus compatible La and Eu showing Rayleigh fractionation models for the Obsidian Creek member rhyolites. Calculated trends show fractional crystallization of parent Willow Park magma (WP-1) in 10% increments. Evolved daughter magmas AS-1, LD-1, PH-5, and GH-2 show good fit (within error) to fractionation trends. Rhyolite and dacite components (GRM-1 and GLM-2) from mingled lavas are not within 2a uncertainty of the fractionation trends.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 2 a e rro r i

60 t Q. Q. 50 £ 1- 40 GH-2

20 GRM-1 WP-1 10 GLM-2

0 140 160

8.50 2 a e rro r "

7.50 -

70 % / ^ H - 2 6.50 - E Q. 5.50 RS 5 0 % / ^ 4.50 - irP H -5

y ^ L D - 1 3.50 - GRM-1 4W W P-1 2.50 - AS-1 •GLM-2 1.50 25 35 45 55 65 75 85 95 105 115 Nb (ppm)

Figure 8.03. Plots of Nb versus incompatible HFSE's Th and Ta showing Rayleigh fractionation models for the Obsidian Creek member rhyolites. Calculated trends show fractional crystallization of parent Willow Park magma (WP-1) in 10% increments. Evolved daughter magmas AS-1, LD-1, PH-5, and GH-2 show good fit (within error) to fractionation trends. Rhyolite GRM-1 and dacite GLM-2 from mingled lavas are not within 2a uncertainty and plot off of the fractionation trends.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 a error ffi 75

70 .GR-2 Q. Q. 65 60 00-5 CS-1

55 60%

50 2 o error h 13 60%

12 r CS-1 OC-5 §: 1 1

1 0 GR-2 9

8 7 2 a error h 0.8

Û. 0.6 CL GR-2

OC-5 0.2 CS-1 60%

0 10 20 30 40 50 60 Nb (ppm)

Figure 8.04. Nb versus LREE's La, Sm, and Eu Rayleigh fractionation models for the Roaring Mountain member rhyolites. Calculated trends show fractional crystallization of parent Gibbon River flow (GR-2) in 10% increments. Evolved daughter magmas Obsidian Cliff (OC-5) and Crystal Spring (CS-1) show good fit on fractionation trends. Error bars display uncertainties at 2 a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 a error h 5 0 ^

OC-5 Q. Q.

GR-2

2 a error h 60%

CS-1 00-5 CL Q. GR-2

2 cj error 50% 3.5 CS-1 OC-5 CL 5 2.5 -

GR-2

0 1 0 2 0 3040 50 60 Nb (ppm)

Figure 8.05. Nb versus HREE Yb, and HFSE's Hf and Ta Rayleigh fractionation models for the Roaring Mountain member rhyolites. Calculated trends show fractional crystallization of parent Gibbon River flow (GR-2) in 10% increments. Evolved daughter magmas Obsidian Cliff (OC-5) and Crystal Spring (CS-1) show good fit on fractionation trends. Error bars display uncertainties at 2a.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 a error h 60% 260

I 220 CS-1 a OC-5 % 180 DU 140 GR-2

100 2 a error I-1 GR-2 250

E 200-

S 150- (Q DO 100 -

50 - OC-5 60% CS-

0 10 20 3040 50 60 Nb (ppm)

Figure 8.06. Plots of Nb versus LIL's Rb and Ba showing Rayleigh fractionation models for the Roaring Mountain member rhyolites. Calculated fractionation trends show fractional crystallization of parent Gibbon River magma (GR-2) in 10% increments. Evolved daughter magmas Obsidian Cliff (OC-5) and Crystal Spring (CS-1) show good fit on fractionation trends. Error bars display uncertainties at 2 a.

1 0 0

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 -

CO SLF-2 GRM-1 a GLM-2 GRM-1

SLF-2 GRM-1 a GLM-2 GRM-1

12 - SLF-2

10 - GRM-1 a

u_ GLM-2

GRM-1

0.2 - SLF-2 GRM-1 a GLM-2 GRM-1

12 - SLF-2

GRM-1 a

GLM-2

GRM-1

40 50 60 70 80 Si0 2

Figure 8.07. S1 0 2 (wt. %) versus major elements (AI2 O3 , Ti0 2 , FeO, MnO, and CaO, in wt. %) mixing models. Plotted are rhyolite end-member (GRM-1) from the Gardner River mingled lava, and a representative basalt end-member (SLF-2) from the Swan Lake Flat basalt. Basaltic-andesite (GRM-1 a) from the GRM-1 host rhyolite, and dacite host lava (GLM-2) from the Grizzly Lake flow (plotted above) can be modeled using ~ 70% SLF-2 with ~ 30% GRM-1, and ~ 40% SLF- 2 with ~ 60% GRM-1 respectively. Basalt data from Bennett (in progress).

101

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SLF-2 GRM-1 a D) GLM-2

GRM-1

GRM-1 GLM-2 GRM-1 a SLF-2

GRM-1 cs SLF-2 GRM-1 a GLM-2

SLF-2 o ’ r 0.2- GRM-la GLM-2

GRM-1

4050 60 70 80 Si0 2

Figure 8.08. S1 0 2 (wt. %) versus major element (MgO, K2 O, Na2 Û, and P 2 O5 in wt. %) mixing models. End members are rhyolite (GRM-1) from the Gardner River mingled lava, and a representative Swan Lake Flat basalt (SLF-2). Basaltic- andesite enclave (GRM-1 a) from rhyolite host GRM-1, and dacite host lava (GLM-2) from the Grizzly Lake flow can be modeled using -70% SLF-2 with -30% GRM-1, and -40% SLF-2 with -60% GRM-1, respectively. Basalt data from Bennett (in progress).

102

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17- co 16- SLF-1 N 15 - AS-13

^ 14- AS-1 13-

SLF-1 2 . 0 - AS-13

AS-1 0.0

10. 0 - SLF-1

8.0 - AS-13

u. 4.0 -

2. 0 - AS-1 0.0

0.3- SLF-1 0.2 - AS-13

AS-1

0.0

12. 0 - SLF-1

8. 0 -

4.0- AS-1 0.0 0 74 5060 7040 80

Si0 2

Figure 8.09. Si0 2 (wt. %) versus major element (AI2 O3 , Ti0 2 , FeO, MnO, and CaO in wt.%) mixing models. End members are rhyolite (AS-1) from the Appolinaris Spring mingled lava, and a representative Swan Lake Flat basalt (SLF-1). Basaltic-andesite enclave (AS-la) from rhyolite host AS-lean be modeled using -80% SLF-1 and -20% AS-1. Basalt data from Bermett (in progress).

103

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10. 0 - SLF-1 8.0 - O AS-13 |> 6 . 0 - 4.0 -

2 . 0 - AS-1 0.0

AS-1 O ^ 4.0-

AS-13 2 . 0 - SLF-1 0.0

4.0 - AS-1 O 3.0- m(M AS-13 Z 2.0- SLF-1

0.0

0.3 - SLF-1 lO AS-13 ° 0.2 - Q.

AS-1 0. 0-1 40 50 60 70 80 Si02

Figure 8.10. S1O2 (wt.%) versus major element (MgO, K2 O, Na2 0 , and P 2 O5 in wt.%) mixing models. End members are rhyolite (AS-1) from the Appolinaris Spring mingled lava, and a representative Swan Lake Flat basalt (SLF-1). Basaltic-andesite enclave (AS-la) from the Appolinaris Spring mingled lava can be modeled using —80% SLF-1 and -20% AS-1. Basalt data from Bennett (in progress).

104

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100

♦ GRM-1 A GRM-1a # SLF-2

m l -

J _ l I I I 1 I I 1 I I I 1 I I I I I I I 1 L CsRbBaTh U Nb K LaCePbPrSr P NdZrSrrEuTiDy Y YbLu 10 E(b) GRM-1a

00 g o o 1 a:

1 CsRbBaTh U Nb K LaCePbPr Sr P NdZrSrÆuTi Dy Y YbLu

Figure 8.11. Sun and McDonough (1989) OIB spider diagram (a) and trace element mixing model (b). Trace element concentrations are displayed (a) for Gardner River rhyolite (GRM-1), Swan Lake Flat basalt (SLF-2), and basaltic- andesite enclave (GRM-1 a). The trace element mixing model (b) shows that basaltic-andesite enclave (GRM-1 a) can be produced by using end members SLF-2 (75%) and GRM-1 (25%). Basalt data from Bennett (in progress).

105

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 F I I I I I I I I I I I I I I I I I I I I I r= ♦ GRM-1 GLM-2 SLF-2

.01 I I I l _ J I I I I I I I I 1— 1 I I I L_J L CsRbBaTh U Nb K LaCePbPr Sr P NdZrSrrEu Ti Dy Y YbLu

10 GLM-2

/ Model

CO O Ü 1 o cr

1 CsRbBaTh UNbK LaCePbPrSr PNdZrSrrEuTiDyYYbLu

Figure 8.12. Sun and McDonough (1989) OIB spider diagram (a) and trace element mixing model (b). Trace element concentrations are displayed (a) for Gardner River rhyolite (GRM-1), Swan Lake Flat basalt (SLF-1), and Grizzly Lake dacite (GLM-2). The trace element mixing model (b) shows that GLM-2 can be produced using end members GRM-1 (60%) with SLF-1 (40%). Basalt data from Bennett (in progress).

106

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 ♦ AS-1 AS-1 a SLF-1

J 1 I I I I I I I I I I I I I I I I I I I L CsRbBaTh UNbK LaCePbPrSr P NdZrSrrEuTiDy Y YbLu

♦ AS-1 jjk AS-1 a 100 SLF-1

■o c o SI y Ü o 10 CL

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 8.13. Sun and McDonough (1989) OIB spider diagram (a) displaying trace element concentrations for end-members (AS-1 and SLF-1) and basaltic-andesite enclave (AS-la). Sun and McDonough (1989) REE diagram (b) shows enrichment of AS-la relative to its rhyolite host (AS-1) of the Appolinaris Spring mingled lava. SLF-1 is plotted to show that mixing between AS-1 and SLF-1 cannot produce the composition of AS-la. Basalt data from Bennett (in progress).

107

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 0 w

* xenolith from SLF basalt flow À AS-2a # SLF-1

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

100

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 8.14. Sun and McDonough (1989) REE plot (a) and trace element mixing model (b). Trace element concentrations are displayed (a) for a granitic xenolith from the Swan Lake Flat basalt, a basaltic andésite enclave (AS-2a) from the Appolinaris Spring mingled lava, and Swan Lake Flat basalt SLF-1. Trace element mixing model (b) shows that AS-2a can be produced using the granitic xenolith (60 %) and SLF-1 (40 %). Basalt data from Bennett (in progress).

108

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SLF-14 SLF-1

SLF-2 ■ SLF-5

"Nd GRM-la

-10 GLM-2 -12 GRM-1

-14 0.705 0.706 0.707 0.708 0.709 0.710 0.711 (87sr/86sr)j

Figure 8.15. (^^Sr/^^Sr); versus epsilon Nd mixing models to produce basaltic- andesite enclave (GRM-la) and dacite (GLM-2). Swan Lake Flat basalts (SLF-1, SLF-2, SLF-5, and SLF-14) were mixed with Gardner River rhyolite (GRM-1) to produce four mixing lines. GRM-la can best be modeled using -75% SLF-2 with -25% GRM-1. GLM-2 can be modeled using SLF-1, SLF-2, and SLF-14. Approximately 30% Swan Lake Flat basalt would need to mix with -70% GRM- 1 to produce GLM-2. Basalt data from Bennett (in progress).

109

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SLF-14 0 # SLF-1

-4

-6 'Nd -8

-10

AS-1 -12

-14 0.704 0.706 0.708 0.71 0.712 0.714 (87sr/86sr)i

Figure 8.16. (87Sr/86Sr)i versus mixing models to produce basaltic-andesite enclave (AS-la) from the Appolinaris Spring mingled lava. End members used were Swan Lake Flat basalt (SLF-land SLF-14) and Appolinaris Spring rhyolite (AS-1). Approximately 35% SLF-1 or SLF-14 would need to mix with -65% AS-1 to produce an isotopic composition similar to AS-la. Basalt data from Bennett (in progress).

110

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15.7

GLM-2 15.6 GRM-1 jQ D. SLF-5 G R M -la s 15.5 S 0 . SLF-2 CMÈ 15.4

15.3

15.2 17.15 17.2 17.25 17.3 17.35 17.4 17.45 17.5 206pb/204pb

Figure 8.17 . 206p^j/204pjj versus mixing models to produce basaltic- andesite enclave (GRM-la) and dacite (GLM-2) from respective Gardner River and Grizzly Lake mingled lavas. Swan Lake Flat basalts (SLF-2 and SLF-5) were mixed with Gardner River rhyolite (GRM-1) to produce two models. Both GRM- la and GLM-2 are best modeled using -75-80% SLF-5 and -20-25% GRM-1. Basalt data from Bennett (in progress).

I l l

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38.4 ■

38.3

38.2- SLF-5 GLM-2 GRM-1 £ 38.1 4 38.0 - GRM-1 IQ. 00 37.9- SLF-2 37.8-

37.7-

37.6-

37.5 17.15 17.2 17.25 17.3 17.35 17.4 17.45 206pb/204pb

Figure 8.18. versus ^^^Pb/^^'^Pb mixing models to produce basaltic- andesite enclave (GRM-la) and dacite (GLM-2) from respective Gardner River and Grizzly Lake mingled lavas. Swan Lake Flat basalts (SLF-2 and SLF-5) were mixed with Gardner River rhyolite (GRM-1) to produce two models. GRM-la is best modeled using -90% SLF-2 and -10% GRM-1. GLM-2 is best modeled using -80% SLF-5 and -20% GRM-1. Basalt data from Bennett (in progress).

112

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15.9

15.8 AS-la 15.7

£ 15.6 AS-1 5 SLF-1 B 15.5 CL SLF-14 6 15.4 CM 15.3

15.2

15.1

15.0 17.45 17.5 17.55 17.6 17.65 17.7 17.75 17.8 206pb/204pb

Figure 8.19. versus ^‘’^Pb/^^^Pb mixing models to produce basaltic- andesite enclave AS-la from the Appolinaris Spring mingled lava. Swan Lake Flat basalts (SLF-1 and SLF-14) were used with Appolinaris Spring rhyolite AS-1) to produce two models. AS-la is not successfully modeled since it is more radiogenic in 2®^Pb/^®'*Pb than the basalt samples SLF-1 and SLF-14. Basalt data from Bennett (in progress).

113

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39.3

39.1

38.9 JQ CL 38.7 S SLF-1 s 38.5 Q. 00 SLF-14 No 38.3 AS-1 38.1

37.9

37.7

37.5 17.45 17.5 17.55 17.6 17.65 17.7 17.75 17.8 17.85 206pb/204pb

Figure 8.20. ^^Pb/^^Pb versus ^°®Pb/^'’'*Pb mixing models to produce basaltic- andesite enclave AS-la from the Appolinaris Spring mingled lava. Swan Lake Flat basalts (SLF-1 and SLF-14) were used with Appolinaris Spring rhyolite AS-1 to produce two models. These models are more successful than the preceding (Figure 8.21) and show that AS-la can approximately be modeled using > 95% SLF-1 or SLF-14 with AS-1. However, AS-la plots off mixing line by more than ± 2ct. Basalt data from Bennett (in progress).

114

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CHAPTER 9

DISCUSSION

Origin of the Extracaldera Rhyolites

Results from this study are used to (1) test the argument from previous work that the

extracaldera rhyolites are not related to the main sub-ealdera silicic magma system, ( 2 )

constrain crustal and mantle sources of the extracaldera rhyolites and mingled lavas, and

(3) identify possible cogeneric relationships among the extracaldera units.

Isotopic Relation to the Main Subcaldera Reservoir Rhvolites

The hypothesis that the extracaldera rhyolites are isotopically and chemically distinct

from the voluminous caldera-related rhyolites is based on previous 6*^0 (Friedman et al.,

1974, Hildreth et al., 1984; Bindeman and Valley, 2000, 2001; Bindeman et ah, 2001)

and Nd, Sr, and Pb (Hildreth et ah, 1991; Doe et ah, 1982) isotopic data. New Nd, Sr, and

Pb isotopic data presented herein support this hypothesis.

The earliest intracaldera rhyolites of the third volcanic cycle show shifts in Nd, Sr,

Pb, and O isotopic ratios thought to be caused by assimilation of roof rocks and/or

interaction with 6 * ^ 0 depleted hydrothermal brines during collapse and resurgence

(Friedman et ah, 1974; Doe et ah, 1982; Bindeman and Valley, 2000, 2001; Bindeman et

ah, 2001). These early postcollapse rhyolites are the most radiogenic in Pb and Sr and

show the greatest depletion in (+1 to +3) (Hildreth et ah, 1984). Successively

younger intracaldera rhyolites exhibit a recovery toward precaldera values marked by

115

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rapid ô**0 enrichment up to +4 to + 6 (Hildreth et al., 1984). These ô'*0 depletion-

enrichment cycles are common primarily to the and 3"^"^ cycle, but are not observed in

the several extracaldera rhyolites analyzed by Friedman et al. (1974) and Hildreth et al.

(1984). The extraealdera rhyolite 0**0 measurements are higher at +6.5 to +7.0

(Friedman et al., 1974; Hildreth et al., 1984) and +5 to + 8 (Bindeman and Valley, 2001)

than any basalts, or most of the caldera related rhyolites, which suggests that the

extraealdera rhyolites were derived from sources independent from the main subcaldera

magma system (Hildreth et al., 1991).

New isotopie results presented herein combined with the previous results of Doe et al.

(1982) and Hildreth et al. (1991) show that the extraealdera lavas are also distinct from

the main subcaldera related rhyolites in their Nd, Sr and Pb isotopie compositions. The

earliest post-collapse rhyolites of both the first and third cycle record apparent

contamination events where *’Sr/*^Sri, ^°^Pb/^°'*Pb and ^®’Pb/^**'*Pb values shift to more

radiogenic compositions and ^‘**Pb/^°'*Pb values shift to slightly less radiogenic

compositions. After these collapse-related contamination events, *^Sr/*^Sr, values recover

to values typical of the main subcaldera reservoir (-0.708 to 0.710), whereas Pb isotope

compositions only partially recover. While the extraealdera rhyolites show considerable

overlap in ^**^Pb/^**'*Pb, ^'*’Pb/^'*"*Pb, and ^***Pb/^**‘^Pb values with the main reservoir

rhyolites, results from this study show that they define a distinct group (excluding daeite

from the Grizzly Lake mingled lava, and geochemically distinct rhyolites from the

Roaring Mountain member) with higher *’Sr/*^Srj values that range from 0.709 to 0.714.

Extraealdera rhyolite Gw values (-12.1 and -13.1) (Table 7.01, Figure 8.01) are also

distinct from the majority of intracaldera rhyolites ( -7.8 to -9.8) (Hildreth et al., 1991).

116

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. These data together with the previously reported 6**0 data require that the extraealdera

rhyolites were derived from a source distinct from that of the main reservoir rhyolites.

Isotonic Relation to Yellowstone Basalts

The isotopie data presented in this study also indicate that the extraealdera rhyolites

are not direct fractional crystallization products of Yellowstone basalts. Hildreth et al.

(1991) has shown that the basalts associated with all three volcanic cycles are less

radiogenic in *’Sr/*^Sri (~ 0.705 - 0.709) and are higher in SNd (— 2 to - 8 ) than both the

main reservoir and the extraealdera rhyolites. The absence of intermediate compositions

and the fact that the extraealdera rhyolites have lower SNd and higher *^Sr/*^Sri values than

Yellowstone basalts argues against direct derivation by fractionation of basalt.

Potential Crustal Source Rocks

Previous studies on regional igneous rocks and associated xenoliths emplaced through

the Archean (~ 3.1 - 2.8 Ga; Leeman et al., 1985) basement of the Wyoming Province

report data that constrain the isotopie composition of the lower crust. Data are available

for Arehean amphibolites from the Beartooth Mountains, Montana (Wooden and Muller,

1988), on Precambrian granitic gneiss from Lamar Canyon near Yellowstone (Doe et al.,

1982), on granulite-facies crustal xenoliths from the Snake River Plain, Idaho-Wyoming

(Leeman et al., 1985; Spencer-Kilgore and Craters of the Moon localities), and on alkalic

and sub-alkalic intrusive rocks as well as homblendite and gneissie crustal xenoliths from

the Crazy Mountains, Montana (Dudas et al., 1987).

Northeast of the Yellowstone caldera complex in the eastern Beartooth Mountains a

compilation of studies (Henry et al., 1982; Mueller et al., 1982, 1985; Wooden et al.,

1982; and Wooden and Mueller, 1988) report *^Sr/*^Sr = 0.712 to 0.783 for granitoids

117

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and 0.712 to 0.936 for metamorphic rocks; SNd values = -36 to -40 for granitoids and as

low as -50 for metamorphic rocks. Doe et al. (1982) report an *^Sr/*^Sr = 0.7427 from

Precambrian granitic gneiss of Lamar Canyon, an area of basement outcrop nearest to

Yellowstone. Southwest of the caldera complex, plutonic xenoliths carried in Quaternary

basalts of the Snake River Plain have ^'^Sr/^^Sr that range widely from 0.702 (related to

Rb depletion accompanying granulite facies metamorphism) to greater than 0.83 (related

to metasedimentary compositions) and SNd = -23 to -35 (Leeman et al., 1985). North of

the Beartooth Range in the Crazy Mountains a gneissie xenolith was analyzed with an

*’Sr/*^Sr = 0.7097 and Ewd = -34 (Dudas et al., 1987).

If the extraealdera magmas were generated directly from partial melts of the crust

similar to these plutonic and metamorphic Archean rocks then they would be expected to

exhibit very low SNd (< -2 0 ) and may show a range of *’Sr/*^Sr due to the diversity of

crustal reservoirs, or to Rb depletion which accompanies granulite metamorphism

(Leeman et ah, 1985). Because the extraealdera rhyolites exhibit significantly higher G^d

(-12.1 to -13.1) and intermediate *’Sr/*^Sri between 0.709 - 0.714, it is improbable that

they were derived directly from partial melts of an Archean source. Also, these rhyolites

cannot represent primary crustal melts since they have very low Sr contents (1-32 ppm),

which require extensive feldspar fractionation (Mahood and Halliday, 1988).

Hvbrid Source End Members

If the magmas parental to the extraealdera rhyolites were not derived directly from a

mantle or crustal source as suggested by the isotopie evidence, then the source is likely a

hybrid of mantle and Archean crustal rocks of the Wyoming Province. Additionally, the

118

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. presence of crustal xenoliths in many basalt flows from the Snake River Plain (Leeman

et. al, 1985) provides evidence for magma-crust interactions in this region.

The results from the preferred hybrid source models (2, 3, and 5) presented in Chapter

8 (Figure 8.01) indicate that the magmas parental to the extraealdera rhyolites require a

basalt contribution between 40 - 70%. Parental rhyolites were likely produced when

basaltic magma evolved to Swan Lake Flat compositions through fractionation and

possibly assimilation, and then stalled in the lower crust promoting melting of country

rock with compositions similar to the Archean granulite xenoliths of the Snake River

Plain area. Mixing of the basaltic and crustal melts then produced a hybridized

intermediate composition magma that ascended to upper crustal levels and differentiated

to high-silica rhyolite compositions. An alternative model is that the hybridized magma

crystallized and was later remelted by renewed basalt intrusion.

Evolution of the Extraealdera Rhyolites

There are two end member models to explain the extraealdera rhyolites in which they

are products of ( 1 ) a series of isolated, independent magma batches that are chemically

distinct and short-lived, or ( 2 ) a single, long-lived magma system that evolved over time.

The geochronologic, pétrographie, and geochemical data favor the second model, and

additionally suggest that the extraealdera rhyolites represent either a single evolving

magma system, or two separate evolving magma batches. Thus, the following sections

discuss the evolution of the extraealdera rhyolite magma system(s) first by examining

initial pétrographie and geochemical relationships constrained by geochronology, and

then by focusing on fractional crystallization and mixing models.

119

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Eruptive Periodicity

The extraealdera rhyolites define two major eruptive periods separated by an ~75 ka

quiescent interval (Figure 9.01). Unrelated to these events are the older Riverside (RF-2)

and Cougar Creek (CC-7) rhyolites, which erupted -170 ka apart at 526 ka and 358 ka,

respectively. The first interval of activity includes six domes/flows that erupt -10-40 ka

apart over the 326-208 ka interval, and the second episode includes four rhyolites that

erupt -12-26 ka apart over the 134-80 ka interval.

Recent "***Ar/^^Ar ages (Bennett, in progress) combined with a previous K/Ar age

(Obradovich, 1992) show that the spatially associated Swan Lake Flat and Osprey basalts

erupted from 350 - 209 ka (Figure 9.01). These ages indicate that eruption of basalt in the

Norris-Mammoth corridor began immediately preceding establishment of the

extraealdera rhyolite magma system (see below) and ceased around 209 ka, whereas

rhyolite domes and flows continued erupting until 80 ka. The extraealdera mingled lavas

(316-263 ka) with basaltic-andesite enclaves were erupting during the first half of the 350

- 209 ka interval of basaltic volcanism.

Evolution of the Extraealdera Magma Svstem

FXL o f the Obsidian Creek Rhyolites

The geochemical relationships presented in this study are consistent with the

interpretation that the Willow Park rhyolite (WP-1) is parental to the Landmark (LD-1),

Paintpot Hill (PH-5), and Gibbon Hill (GH-2) evolved rhyolites of the Obsidian Creek

member. Production of the evolved daughter compositions are modeled by -60 - 70%

crystallization of the parent magma (Figures 8.02 and 8.03) using trace element

concentrations and the modal phenocryst phases and abundances of WP-1 (Table 5.01).

120

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LREE depletion with time indicates allanite fractionation and Eu depletion records

crystallization of feldspars. The extremely low Sr concentrations (< 2 ppm) of LD-1, PH-

5, and GH-2 rhyolites (Table 7.01) may further indicate extensive feldspar fractionation.

Although Willow Park rhyolite was chosen as the parental magma based on its age

(Figure 9.01), trace element data together with spatial and temporal relationships suggest

that any of the three earliest erupted rhyolites (Willow Park at 326 ka, Appolinaris Spring

at 316 ka, and Gardner River at 301 ka) could represent a parental composition.

REE plots show that these older, least evolved rhyolites have steep LREE patterns

that overlap whereas successively erupted rhyolites (LD-1, PH-5, and GH-5) have

distinctly flat LREE patterns and show a decreasing trend with time (Figure 7.06). Eu

anomalies and HREE patterns are consistent in showing that WP-1 and AS-1 are

indistinguishable, yet GRM-1 is comparatively more evolved as indicated by a deeper Eu

anomaly and enrichment of HREE’s. Spatially, the Willow Park and Appolinaris Spring

domes erupted < 2 km apart along a linear trend most likely controlled by north striking

normal faults (Figure 2.02), whereas the Gardner River flow erupted ~ 4 km to the west.

'***Ar/^^Ar ages at the 2 a level of confidence indicate that as few as 2 ka may separate

eruption of the Willow Park (326 ± 4 ka) and Appolinaris Spring (316 ± 4 ka) domes and

5 ka may separate eruption of the Appolinaris Spring dome from the Gardner River flow

(301 ± 6 ka).

The possibility that WP-1 and AS-1 represent eruption of the same magma is also

supported by similar modal abundances of phenocrysts and the fact that both are the only

rhyolites of the Obsidian Creek member that contain clinopyroxene (Table 5.01).

Although clinopyroxenes can crystallize as primary phenocrysts in rhyolitic magmas, it is

121

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. also possible that they are xenocrysts derived from Swan Lake Flat basalt. However,

compositional data from phenocryst phases necessary to constrain origin of the

clinopyroxene was not obtained in this study. A Swan Lake Flat origin would indicate

that the Willow Park rhyolite underwent a mingling event that potentially went unnoticed

due to extensive soil cover and limited outcrop exposure, whereas a rhyolite origin would

satisfy the lack of mafic inclusions in sample WP-1. The Gardner River rhyolite (GRM-

1) is aphyric and therefore petrographically distinct from WP-1 and AS-1.

The Appolinaris Spring rhyolite thus likely represents eruption of the Willow Park

magma from an adjacent vent within a short time interval. The greater Eu anomaly,

evolved HREE pattern, and significantly younger age of GRM-1 compared to WP-1 and

AS-1 suggests it represents a fractional crystallization product of parental Willow

Park/Appolinaris Spring magma.

FXL of the Roaring Mountain Rhyolites

Fractional crystallization models (Figures 8.04, 8.05, and 8.06) suggest that the

Gibbon River rhyolite (GR-2, 118 ka) is parental to the more evolved Obsidian Cliff

(OC-5, 106 ka) and Crystal Spring (CS-1, 80 ka) rhyolites. Production of the evolved

daughters are modeled by -45 - 50% crystallization of parental Gibbon River magma

(Figures 8.04, 8.05, and 8.06).

The aphyric Crystal Spring and Obsidian Cliff flows, which erupted only -0.5 km

apart, may represent eruption of the same magma through separate conduits based on

their indistinguishable trace element chemistry (Figure 7.07). Although the eruptive ages

of these flows are distinct (80 ± 2 ka. Crystal Spring and 106 ± 1 ka. Obsidian Cliff)

122

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (Table 6.01), they are based on analysis of glass and could be inaccurate due to problems

associated with dating glass (Chapter 6 ).

Magma Mingling

Mingled Lavas of the Obsidian Creek Member

Three mingled lavas (one newly discovered) were erupted early in the lifespan of the

Obsidian Creek magma system in the northern part of the study area (Figure 2.02); the

Appolinaris Spring dome at 316 ka, the Gardner River flow at 301 ka, and Grizzly Lake

flow at 263 ka.

The Appolinaris Spring dome is a newly discovered mingled lava that documents

input of mafic magma at the onset of extraealdera volcanism. Mingling of Appolinaris

Spring rhyolite with a mafic magma apparently did not affect its original major or trace

element concentrations since AS-1 is a high-silica rhyolite that overlaps (at 2a) with WP-

1 on fractionation trends (Figures 8.02 and 8.03) and has an indistinguishable REE

signature from WP-1 as discussed above. AS-1 and WP-1 appear to represent the same

rhyolitic magma, however mafic enclaves were not seen in the Willow Park rhyolite.

This suggests that intruding basalt only interacted with an isolated region of silicic

magma to produce the Appolinaris Spring mingled lava, and did not affect the magma

body on a system wide scale. Alternatively, residual magma from the Willow Park

eruption may have mingled with basalt to produce the Appolinaris Spring mingled lava.

Mixing models suggest that basaltic-andesite enclaves from the Appolinaris Spring

mingled lava represent hybridized magma. The most reasonable mixing end members are

the Appolinaris Spring rhyolite (AS-1) and the contemporaneous Swan Lake Flat basalt

based on temporal and spatial relations. Major element, trace element, and isotopic

123

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mixing models, however, are not all consistent in showing that the enclaves were

produced from hybridization. Furthermore, major (Figure 7.01) and trace element (Figure

7.08, top) analyses of two analyzed basaltic-andesite enclaves (AS-la and AS-2a) suggest

that they are heterogeneous.

Although the major element compositions of AS-la can be consistently modeled by

mixing -80 % Swan Lake Flat basalt (AS-2a is similarly modeled with - 90%) with AS-1

rhyolite (Figures 8.09 and 8.10), these proportions are not consistent with isotopie mixing

models (Figures 8.16 and 8.20), which require -35% (*^Sr/*^Sr, vs. SNd model) and >95%

(206pb/204pb y g model) basalt (isotopie data was not available to model AS-

2a). Additionally, the uncharacteristic enriched REE signatures of AS-la and AS-2a

(Figure 7.08, top) relative to AS-1 are impossible to model by mixing the two end

members.

However, a granitic xenolith from the Swan Lake Flat basalt (Bennett, in progress)

does have REE concentrations similar to what is required to model the basaltic-andesite

enclaves. Figure 8.14b shows that mixing 40% Swan Lake Flat basalt with the granitic

xenolith can produce AS-2a, however it is unknown whether major element and isotopie

mixing would be consistent with the trace element model since these data are not

available. Thus, the enclaves likely represent a hybrid of basaltic and silicic magma,

however, the silicic end-member cannot be confidently constrained. An alternate model

in which the basaltic-andesite enclaves are direct fractionates of the Swan Lake Flat

basalt is not consistent with their distinct isotopic compositions (Figure 8.16).

The rhyolite host (GRM-1) of the Gardner River mingled lava plots off most

fractionation models (Figures 8.02 and 8.03) since its LREE element concentrations are

124

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. slightly less evolved than its parent (WP-1). However, the Eu fractionation model (Figure

8.02) suggests that-10% fractionation of WP-1 can produce the Eu concentrations of

GRM-1, which indicates feldspar crystallization. As previously discussed, LREE

depletions with time among the Obsidian Creek rhyolites indieate allanite fractionation

(Figure 8.02). That GRM-1 is enriched in LREE’s and does not fit on LREE fractionation

models may simply indicate that allanite was not a fractionating phase in the Willow Park

magma until sometime after the Gardner River flow erupted. Thus, GRM-1 may

represent - 10% fractionation of Willow Park magma. Alternatively, the possibility that

the original trace element eoncentrations of GRM-1 were slightly modified from

mingling with a mafic magma may be why the fractionation models are only partly

successful.

The basaltic-andesite enclaves (GRM-1 a) within this mingled lava are interpreted to

represent a hybrid magma derived from mixing between Swan Lake Flat basalt and the

host rhyolite GRM-1. Major element (Figures 8.07 and 8.08), trace element (Figure 8.11),

and isotopie (Figures 8.15 and 8.17) mixing models are remarkably consistent in showing

that the enclaves can be produced by mixing -70% basalt with the host rhyolite. The

crenulated silicic-mafie interfaces (Figure 5.03) and exchanged phenocrysts of reaction-

rimmed quartz in enclaves indicate that the mafie and silicic components were liquid at

the time of mingling. Disequilibrium textures such as resorbed sanidine phenocrysts with

reaction rims (Figure 5.04) in rhyolitic glass, and strongly zoned plagioclase and resorbed

pyroxenes in mafic enclaves is eonsistent with liquid state interaetion between these

magmas.

125

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Major element (Figures 8.07 and 8.08), trace element (Figure 8.12) and isotopie

(Figures 8.15, 8.17, and 8.18) mixing models consistently show that mixing -30-40%

Swan Lake Flat basalt with rhyolite (GRM-1) from the preceding Gardner River flow

eruption can produce the dacite host (GLM-2) of the Grizzly Lake mingled lava. That

GLM-2 diverges farthest from fractionation trends is also consistent with a derivation by

mixing. Hybridization to a dacitic composition must have been restricted to an isolated

region of the magma system since subsequent eruptions of the Landmark, Paintpot Hill,

and Gibbon Hill rhyolites show a continued evolution by fractional crystallization of the

Willow Park magma.

It is concluded in Appendix C that sample GLM-2a, originally thought to represent a

mafic enclave within the Grizzly Lake host lava, instead represents a collected andésite

xenolith of the Washburn volcanics. Thus, the actual mafic clots and enclaves observed

within the Grizzly Lake host dacite were not analyzed. However pétrographie analysis

shows that their mineralogy is similar to the basaltic-andesite enclaves and clots from the

Gardner River and Appolinaris Spring mingled lavas (Table 5.01). The crenulated dacite-

mafic clot interfaces and disaggregated plagioclase and clinopyroxene phenocrysts that

migrate from clots into the dacite glass (Figure 5.05) indicate that the two were liquid at

the time of mingling. The presence of brown glass surrounding mafic clots provides

further evidence that the mafic lava may have quenched against the cooler dacite.

Mingled Lavas o f the Roaring Mountain Member

The Crystal Spring flow is a newly discovered mingled lava in the Roaring Mountain

member that documents evidence of mafic input into the magma system as recently as 80

ka. The chemistry of the mafic magma is unknown since enclaves were not analyzed. A

126

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. thin section of sample CS-1 a reveals the small scale and chaotic nature of inter-layering

between the silicic and mafic components. The crenulated silicic-mafic interfaces

between layers indicate liquid state mingling. The flowbanded nature of the rhyolite

around flattened sub-spherical mafic clots, which likely disaggregated from the mafic

magma and chilled against the cooler rhyolite magma, provides additional evidence for

liquid-liquid interaction. However, the observation that the major and trace element

chemistry of the Crystal Spring rhyolite (CS-1) is indistinguishable from the Obsidian

Cliff flow (OC-5), thought to represent the same magma, indicates that mingling with a

mafic magma did not significantly affect the chemistry of the Crystal Spring rhyolite.

Another observation is that CS-1 is a high-silica rhyolite, which suggests that it did not

hybridize much, if at all.

Transition to the Roaring Mountain Member Rhvolites

The consistent breaks in evolutionary trends between the 326-134 ka and 118-80 ka

rhyolites document an important chemical change to the Obsidian Creek magma system

since trace element concentrations at 118 ka become significantly less evolved (Figures

7.04 and 7.05). This is also observed in REE plots (not shown) in which the least evolved

118 ka Roaring Mountain member rhyolite (GR-2) overlaps with the older, least evolved

Obsidian Creek member rhyolites (WP-1, AS-1 and GRM-1) in LREE’s and HREE’s.

Two main models may explain the chemical and pétrographie transition between the

326-134 ka (Obsidian Creek) and 118-80 ka (Roaring Mountain) rhyolites. First is that

the two intervals of rhyolitic volcanism represent eruption of individual magma batches

from separate magma systems. Alternatively, the Obsidian Creek magma/crystal mush

was recharged either by mixing with a less evolved silicic magma, or by reheating/

127

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. remelting residual erystalline material by influx of hotter magma that did not physically

mix. Recent similar explanations for thermal rejuvenation of a partly solidified magma

body following renewed mafic input have been proposed for silicic to intermediate

magmas at the Soufriere Hills, Montserrat (Zellmer et al., 2003), the San Juan voleanic

field (Bachmann et al., 2002), and the Vinalhaven intrusive complex of coastal Maine

(Wiebe et al., 2004).

The distinct petrography of the 326-134 ka interval rhyolites from the 118-80 ka

interval rhyolites provides additional constraints on this transition. Rhyolite domes and

flows that erupted from 326-134 ka are porphyritic, whereas flows from the 118-80 ka

interval are dominantly aphyric. The transition from a porphyritic to an aphyric texture

suggests that the temperature of the Obsidian Creek magma system was increased after

134 ka and was sustained at higher temperatures for -40 ka. The Crystal Spring mingled

lava, which includes abundant mafic enclaves, is evidenee for an influx of hotter mafic

magma into the magma system as recently as 80 ka. The volume of magrrta erupted also

increased during the 118-80 ka interval as indieated by the Gibbon River and Obsidian

Cliff flows, which are among the largest of the extraealdera rhyolites (Figure 2.02). The

Crystal Spring flow is comparatively small in volume. However it may actually represent

eruption of the Obsidian Cliff magma through a separate conduit, as previously

diseussed.

The transition between the two groups is recorded by the Gibbon Hill dome, which

represents the youngest most evolved Obsidian Creek rhyolite, and the subsequently

erupted Gibbon River flow, which represents the least evolved of the Roaring Mountain

rhyolites. The eomparatively large volume Gibbon River rhyolite erupted along the

128

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. caldera margin and flowed around the Gibbon Hill dome (Figure 2.02). The fact that

these two distinct rhyolites erupted adjacent to eaeh other and only-16 ka apart (Table

6 .0 1 ) strongly suggests that they were derived from the same underlying magma

reservoir. This idea is also consistent with the fact that the Obsidian Creek and Roaring

Mountain member rhyolites cluster in their isotopie compositions. Thus, a model in

which the two were derived from the same magma system is more consistent with these

observations than a model in which a separate magma batch was emplaced over a short

interval of time as an adjacent magma system with isotopic compositions

indistinguishable from the Obsidian Creek magmas.

If derived from the same magma system the apparent up-temperature transition to

aphyric magma and the fact that the Gibbon River rhyolite is distinctly less evolved needs

to be explained. Recharge of the Obsidian Creek magma system by mixing with a less

evolved magma is one possibility, however whether this occurred with a mafic or more

silicic composition is difficult to constrain. Thermal input supplied by intrusion of mafic

magma presumed to be responsible for the rejuvenation and remobilization of the

Obsidian Creek magma body may be represented at the surface by the presence of mafic

enclaves in the Crystal Spring flow.

An alternative model is one in which the crystal mush plus residual magma from the

Gibbon Hill eruption is reheated by intrusion of subjacent hotter magma that did not

physically interact and mix. In this type of model underlying mafic magmas can flux heat

and fluids, which promote melting in the overlying crystal mush (Bachmann et al., 2002)

producing a less evolved magma. The predominantly aphyric nature of the Gibbon River

rhyolite may record reheating of the magma system. The Gibbon River rhyolite, sampled

129

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. in this study near the base at -2300 m and by Christiansen (2001) at -2400 m, is

described in both studies as black obsidian that includes sparse small phenocrysts of

sanidine and quartz. The latest material to erupt (sampled at -2500 m), which appears to

have built up the middle region of the flow (2500-2600 m), is more porphyritic and

contains -13% phenocrysts of sanidine and quartz with lesser amounts of plagioclase and

clinopyroxene (sample GR-2, Table 5.03). This pétrographie transition may represent an

aphyric to porphyritic eruptive sequence as would be expected if the Gibbon River

rhyolite erupted from a zoned magma chamber that contains more aphyric material at the

roof and more crystal rich material in the deep interior. In this case the more volatile-rich,

more evolved, aphyric magma at the roof would have initially erupted, followed by a

progressive tapping of deeper, less evolved, porphyritic magma.

Most magma chambers are traditionally thought to crystallize from their roof inward

and thus, are expected to be hotter at depth. The more aphyric nature of the roof region in

this model may be attributed to higher volatile concentrations, which suppress

crystallization (Bachmann et al., 2002), or to extraction of crystal poor interstitial melt

from a crystal mush (Bachmann and Bergantz, 2004).

Models for Silicic Volcanism

The chemically related extraealdera rhyolites represent eruptions from a high-silica

magma system over an -250 ka period and thus, offer the opportunity to test models for

the generation, storage, and evolution of silicic magma in the upper crust. Proposed

models can be divided into those in which silicic magmas are stored as thermally stable

large chambers for timescales approaching or exceeding 1 Ma (Hildreth, 1981; Halliday

130

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. et al., 1989; Davies et al., 1994; Reid et al., 1997; Davies and Halliday, 1998; and

Heumann et al., 2002), or those in which silicic magmas are generated and erupted

rapidly with short residence timescales of « 1 0 0 ka (Huppert and Sparks, 1988; Mahood,

1990; Wolff and Gardner, 1995; Reid and Coath, 2000; Bindeman and Valley, 2001;

Annen and Sparks, 2002; Bachmann et al., 2002; Woods and Huppert, 2003).

These models have developed from studies at long-lived caldera-forming silicic

magma systems such as the Long Valley Caldera system (California), the Valles Caldera

system (New Mexico), and the Yellowstone Plateau volcanic field (Wyoming). A study

of the generation and storage timescales of pre-emptive magmas at Long Valley by

Halliday et al. (1989) initiated a series of comments and replies (Sparks et al., 1990;

Halliday, 1990; Mahood, 1990) which tested the debated models and have since

encouraged many additional studies (Christensen and Halliday, 1996; Reid et al., 1997;

Reid and Coath, 2000; Heumann et al., 2002). Compared to the Long Valley ealdera

system, the Valles (Wolff et al., 1999, 2002) and Yellowstone (Bindeman and Valley,

2001; Bindeman et al., 2001; and Vasquez and Reid, 2002) ealdera systems have only

recently been the focus of such work.

Magma Residence Timescales

Recent studies have used ^°^Pb/^^*U, ^^**Th/^^*U, and 0**0 isotope data from

crystallizing phases (eg: quartz, feldspar, zircon) in order to constrain timescales of

magma residence and differentiation. None of the above referenced investigations at

Yellowstone have focused specifically on the extraealdera rhyolites, however recent work

(Spell and Nastanski, 2004) is changing this. Vasquez and Reid (2002) previously

131

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. established potential magma residence times up to -150 ka for the extraealdera magma

system based on zircon ages from the Gibbon River flow (118 ka, eruptive age).

This new evidenee for substantial magma residenee times encouraged Spell and

Nastanski (2004) to analyze zircons from four additional rhyolites. zircon ages

from the Willow Park dome (326 ka, eruptive age) range from 280 to 403 ka and from the

Gardner River flow (301 ka, eruptive age) range from 300 to 437 ka. ^^**Th/^^*U zircon

ages from the younger Paintpot Hill dome (208 ka, eruptive age) range from 205 to 225

ka and from the Gibbon Hill dome (134 ka, eruptive age) range from 122 to 180 ka. Four

analyzed spots from the Paintpot Hill zircons yielded seeular equilibrium conditions

(older than -300 ka). ^**^Pb/^*U ages of these spots range from 276 to 311 ka. Results of

this preliminary work suggest magma residenee times from - 1 0 0 ka to within uncertainty

of eruption age. Long residence times are surprising for these rhyolites, which have

average volumes <1 km^ and total only -7 km^ (Hildreth et al., 1990), since small

volumes may be perceived as the products of ephemeral magma batches.

Long residenee times have similarly been established for the Deer Mountain (115 ka,

eruptive age) and South Deadman (-0.6 ka, eruptive age) post-caldera rhyolites at Long

Valley (Reid et al., 1997). Despite their small volume (< 1 km^) and the fact that they

erupted more than 100 ka apart, the majority of Th/ U zireon ages from the two lavas

cluster at 230 ka indicating that they erupted from a common silicic magma reservoir

with residence times > 1 0 0 ka.

Application of End Member Models

Proposed meehanisms to prevent upper crustal magma bodies from cooling below

liquidus temperatures over long periods of time include either new additions of similar

132

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. magma supplied by repeated melting in the crust, or input of mantle-derived basalt

(Hildreth, 1981; Halliday et al., 1989). For example, geochronologic and geochemical

data from the Glass Mountain rhyolites (Long Valley) are interpreted to represent the

existence of a long-lived (300-700 ka), large-volume magma chamber, maintained by

intrusion of mantle-derived basalt (Halliday et al., 1989). This type of model is not

unreasonable for the extraealdera rhyolites, which are distributed across an area -25 km

in length and thus suggest the existence of a laterally extensive magma body. The

longevity of this magma body is constrained by geochronologic data that indicate these

rhyolites erupted over a long time period (-250 ka) and suggest that silicic magmas were

stored on the order of 10"* years (Spell and Nastanski, 2004). Additionally, the

extraealdera rhyolites are interpreted to be genetically related and to have evolved from a

single magma system over their entire eruptive period, and their low Sr contents

(Appendix B, Table 1) are consistent with the Halliday et al. (1989) model for extensive

feldspar fractionation. The presence of mafic enclaves in the rhyolites may be evidence

that intrusion of mafic magma was supplying the heat necessary to sustain a magma body

of this presumed size for long periods of time. However the thermal input necessary to

balance the heat loss by conductive cooling in the upper crust may is argued to be a

geologic problem (Huppert and Sparks, 1988; Sparks et al., 1990). Thus, other models

which seem to explain the above observations without the need to maintain the ideal

temperature eontrol, may be more physically feasible.

Huppert and Sparks (1988) propose that permanent magma chambers need not exist

and that apparent long-lived silicic systems can instead be explained by discrete melting

events. Their model involves the progressive heating of the crust by repeated basaltic

133

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. intrusions, which can erupt at the surface if the crust is initially cold (Figure 9.02a).

Focused intrusions of basalt can progressively heat the interior of the crust to

temperatures that cause melting. Magma batches are generated rapidly at depth and can

either ascend uninterrupted to the surface, or be emplaced at shallow levels where they

could further differentiate to more evolved magma before erupting, or crystallize to form

a pluton (Figure 9.02b). Once extensive partial melting occurs, sufficient volumes of

silicic magma form a density barrier that blocks the ascent of basalt. Ductile crust can

further prohibit basalt intrusions from ascending to the surface due to a lack of brittle

pathways. The continued entrapment of basalt beneath this region aids in the formation of

large volumes of silicic magma that can ascend and cause major caldera-forming

eruptions (Figure 9.02c).

The Huppert and Sparks model (1988) may have some bearing on the early stages of

extraealdera volcanism, however it is inconsistent with the geochemical continuity of the

extraealdera rhyolites over lOO’s of ka’s. Eruption of the Swan Lake Flat basalts prior to

rhyolitic volcanism could potentially represent the initial stage (Figure 9.02a) where

basalt can intrude through the cold crust and reach the surface. Establishment of the

extraealdera magma system may represent the second stage (Figure 9.02b) in which

silicic melt was generated by basalt induced partial melting of a deep crustal reservoir

and ascended to shallow levels. However, the long-term differentiation processes

recorded by the extraealdera rhyolites such as evolution by fractional crystallization,

magma mingling/mixing, and rejuvenation events do not fit this batch-type model since it

suggests that magmas are generated rapidly, rise to the upper crust, and crystallize.

134

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Alternative Models

Bindeman et al. (2001) present an alternative mechanism for the rapid generation of

silicic magma, in which post-collapse intracaldera volcanism at Yellowstone involves

shallow remelting of collapsed wall rock. This model was proposed based on low- 0**0

intracaldera lavas (0.5-0.4 Ma) in which ion microprobe U-Pb ages (2.4-0.5 Ma) and

8**0 analysis of zircons indicate inheritance from pre-caldera volcanic rocks (Bindeman

et al., 2001). However, this model does not apply to the petrogenesis of the extraealdera

rhyolites, which have normal (+5 to +8 ) 8**0 values (Friedman et al., 1974; Hildreth et

al., 1984; Bindeman and Valley, 2001).

Bachmann and Bergantz (2004) present a new model for the origin of high-silica,

crystal poor rhyolites that involves extraction of interstitial melt from magmas of

intermediate composition (andésite to dacite) after they have reached ~ 40-50%

crystallization (Figure 9.03). This model is based on the observation that voluminous

crystal mushes preserved either as batholiths, or erupted as crystal-rich ignimbrites

contain interstitial melts that are high-silica rhyolite in composition (Bachmann et al.,

2002). Bachmann and Bergantz (2004) argue that groups of oogenetic small volume

rhyolites that erupted over long time intervals (-100-500 ka), similar to domes at the

Coso volcanic field (Manley and Bacon, 2000), can be explained as originating from a

common silicic melt that segregated from a large region of crystal mush in the upper

crust. Cogenetic rhyolites erupted over long time intervals imply derivation from a long-

lived reservoir, although their small volumes do not demand the presence of a large

crystal-poor magma body sustained throughout the entire eruptive period (Bachmann and

135

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bergantz, 2004). Eruption of small volume domes and flows can thus be explained as the

result of tapping a rhyolitic melt horizon (Figure 9.03c).

This model can reasonably explain a majority of the observations from the

extraealdera magma system and does not require the sustaining of an entire magma body

at temperatures on the liquidus for lOO’s of ka’s. The hybrid parental magma of the

extraealdera rhyolites was likely intermediate in composition based on mixing models

(Figure 8.01) and isotopic data, and may have been emplaced at shallow levels following

production in the deep crust by basalt intrusion (Huppert and Sparks, 1988). According to

the model of Bachmann and Bergantz (2004) it is not necessary that this entire magma

body underwent extensive differentiation to a high-silica rhyolite composition. Instead,

up to -500 km^ of crystal-poor rhyolite can segregate in 10"*-10^ years from a high

porosity crystal mush. Preliminary and ^*°Th/^^*U zircon ages from the

extraealdera rhyolites suggest residence times up to lO"* years (Vasquez and Reid, 2002;

Spell and Nastanski, 2004). This would be a sufficient amount of time to expel large

volumes of crystal-poor melt into a segregated horizon. A distribution of magma

residence times from > 1 0 0 ka to within uncertainty of eruptive ages suggests that zircons

continuously crystallized from a single, chemically evolving magma. The oldest zircon

age of 437 ka may correspond to an episode of crystal-melt segregation and may

constrain the timing of establishment of the extraealdera rhyolite magma system.

Alternatively, zircons that yield long residence times may be derived from the long-lived

mush in which they started growing (Vasquez and Reid, 2002; Bachmann and Bergantz,

2004). This crystal mush could have been sustained for long timescales due to thermal

input from below. This idea was originally proposed by Mahood (1990) who suggested

136

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. that magmas could thermally oscillate and experience periods of crystallization and

remelting throughout their lifespan as a result of periodic mafic input.

Following the model of Bachmann and Bergantz (2004), the small volume and

crystal-poor nature of the extraealdera rhyolites could be explained as the periodic

tapping of a crystal-poor rhyolite reservoir over -250 ka, and thus can account for the

long-term chemical evolution of the magma system. Eruption of mingled lavas at 316 ka,

301 ka, 263 ka, and again at 80 ka may indicate time intervals of major mafic intrusive

activity.

A Multi-stage Model for the Extraealdera Magma Svstem

Consideration of the above models suggests a multi-stage model for the origin and

evolution of the extraealdera rhyolites. Figure 9.04 summarizes the following details of

this model as a simplified three-stage schematic that begins with establishment of the

magma system after the intermediate hybrid source ascended to upper crustal levels and

was emplaced beneath the Norris-Mammoth corridor as a crystal rich magma.

Stage 1 (Figure 9.04a)

The occurrence of zircon ages up to 437 ka (Spell and Nastanski, 2004) from the

extraealdera rhyolites may record the time when crystal-melt fractionation (Bachmann

and Bergantz, 2004) of the intermediate mush began to produce high-silica rhyolitic

magma. Mafic magmas interacted with this newly established magma system in several

different ways. Some intrusions did not intersect the magma body were thus able to

ascend directly to the surface. This is represented by eruption of the Swan Lake Flat

basalts (shown in yellow. Figure 9.04a) beginning at -350 ka at the northern margin of

extraealdera volcanism. By 326 ka, the rhyolitic magma horizon attained the chemical

137

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and isotopic composition of the extracaldera rhyolites as recorded by the Willow Park

dome, which is interpreted to represent the parental magma for subsequent rhyolites.

Magma mingling and mixing events at 316, 301, and 263 ka may have occurred in

isolated portions of the magma body since system wide fractional crystallization

producing successively more evolved rhyolitic magma continued. Eruptions of mingled

lava were likely triggered when mafic intrusions penetrated the northern region of the

magma body (Figure 9.04a) and interacted with rhyolitic melts.

Stage 2 (Figure 9.04b)

Subsequent eruptions between 263 and 134 ka were controlled by fractional

crystallization of the parental Willow Park magma. This period of activity may represent

system wide cooling and crystallization of the magma reservoir due to a lull in mafic

input. The underlying ponded basalt from previous episodes of mafic intrusion may have

provided the thermal support to sustain the crystal mush at near liquidas temperatures

during this interval. Swan Lake Flat basalts cease erupting after 209 ka, which may be

attributed to either a cessation of mafic input, or to the trapping of basalt beneath the

silicic magma body.

Stage 3 (Figure 9.04c)

After 134 ka the magma system was thermally rejuvenated by renewed input of

mafic, volatile-rich magma that promoted remelting of the residual rhyolitic crystal mush,

producing less evolved aphyric magma. Subsequent eruptions that tapped this reservoir

over the 118-80 ka interval are consistent with evolution by fractional crystallization, and

are characterized by larger volume flows with an aphyric texture as compared to the

smaller volume porphyritic Obsidian Creek rhyolites. Eruption of the Crystal Spring

138

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mingled lava indicates that mafic magma was continuing to flux into the magma system

as recently as 80 ka.

Implications for a 4* Volcanic Cycle

Whether effusive small volume rhyolite eruptions represent the beginning or final

stage of a caldera-forming volcanic cycle (Wolff and Gardner, 1995; Reid et al., 1997)

continues to be a geologic problem. The Yellowstone caldera complex, for example, is

recognized to have produced three major ignimbrite cycles and as a result the early

products of each cycle are commonly destroyed or buried. Because the young

extracaldera rhyolites represent a magma system that is chemically distinct from the

adjacent sub-caldera magma system, they may represent the establishment of a new

magma system that has the potential to build to caldera-forming eruptions.

The significance of the cogenetic extracaldera rhyolites in terms of their petrology,

chemistry, age, and geographic location is that they may represent the onset of a new

cycle of activity due to renewed emplacement of basaltic magma into the crust north of

Yellowstone Caldera. The longevity of this cycle would be controlled by the magnitude

and persistence of mafic intrusive activity in this region. Although this is difficult to

constrain, eruption of the Swan Lake Flat basalts prior to and contemporaneous with the

early evolution of this magma system, and the presence of mafic magma of Swan Lake

Flat affinity in early and late erupted mingled lavas indicates that mafic intrusive activity

has either been ongoing, or has occurred in discrete pulses throughout the ~ 250 ka

lifespan of this system. Additionally, preliminary U-Th-Pb isotope analyses of zircon

indicate that silicic magmas were crystallizing in this area beginning at -450 ka (Spell

139

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and Nastanski, 2004) and continuing to at least 80 ka based on the K/Ar eruptive age of

the Crystal Spring rhyolite. Eruption of these rhyolites within the Norris-Mammoth

corridor thus identifies this region as a potential site for future volcanism including major

explosive eruptions.

The effusive versus explosive nature of the extracaldera rhyolites has implications for

whether the magma system underlying the Norris-Mammoth corridor has the potential to

build to an explosive eruption style. The effusive eruptions of the extracaldera rhyolites

may be explained in several ways. If the volatile content of the magma is high, then the

likelihood of an explosive eruption occurring is much greater than for low volatile

contents. Measurements of volatile contents from melt inclusions in minerals such as

quartz can verify this, however, these data are not available for the extracaldera rhyolites.

Alternatively, the small volume and effusive nature of extracaldera rhyolite eruptions can

be attributed to normal faults periodically intersecting and venting the magma chamber.

As differentiation proceeds, volatiles are prevented from building to high enough levels

to trigger an explosive or caldera-forming eruption. The location of numerous normal

faults in the Norris-Mammoth corridor (Christensen, 2001) and the observation that the

Gibbon River flow vented on the caldera ring fracture and the Obsidian Cliff flow

partially conceals a preexisting normal fault (Christensen, 2001) suggests that faulting

played a role in the location and effusive eruption of these rhyolites. Eruption of the

chemically indistinguishable Willow Park and Appolinaris Spring magmas along two

separate conduits oriented with normal faults is also consistent with this idea. Other

examples that relate faulting to the effusive eruption of small volume, high-silica

rhyolites include a group of 32 chemically related high-silica rhyolite domes at the Coso

140

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. volcanic field (Manley and Bacon, 2000) and the Bearhead rhyolite domes in the Jemez

volcanic field (Justet and Spell, 2001).

A potential argument against the existence of a persistent magma system beneath the

Norris-Mammoth corridor is that seismic studies have thus far failed to image a magma

system in this region. Although low P-wave velocities and P-wave to S-wave velocity

ratios (Vp/Vg) (Miller and Smith, 1999; Husen et ah, 2003) have been detected beneath

the Norris-Mammoth corridor, they are interpreted to represent low-velocity sediments

that fill a deep graben bounded to the west by the Gallatin fault (Figure 2.02). These low

velocity anomalies are also explained as areas that are fluid and/or gas-saturated due to

interaction with associated hydrothermal features in this region (Miller and Smith, 1999).

One reason that a magma body has not been tomographically imaged at deeper levels

may be due to the quality and resolution of the data. Husen et al. (2003), for example,

used local earthquake data collected from nineteen one-component and six three-

component seismometers at Yellowstone. A majority of earthquake hypocenter locations

strongly cluster northwest of Yellowstone Caldera near Hebgen Lake, MT, with a smaller

group restricted to the southern end of the Norris-Mammoth corridor, and a sparse

number within the caldera. S-wave and P-wave velocities being recorded in the Norris-

Mammoth corridor are limited since only one three-component station (Norris) and one

one-component station (Mammoth) are spatially distributed. Therefore the lack of

earthquake events and seismic stations in this region yields poor source (earthquake) to

receiver (station) ray path coverage across the Norris-Mammoth corridor. Thus, it is

possible that a Norris-Mammoth magma body has not been resolved due to poor

coverage and resolution of the data. For example, Vp anomalies of small size (15x15 km)

141

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. have been reliably resolved northwest of the caldera where ray path eoverage is high,

whereas only Vp anomalies of large size have been reliably imaged inside the caldera

where ray paths are lacking.

Hydrothermal activity in the Norris-Mammoth corridor is additional evidence that a

magma body may exist beneath this region. Hydrothermal activity in Yellowstone is

structurally controlled and most features within the caldera and along the caldera ring

fracture zone have largely been attributed to cooling of the sub-caldera magma system. A

large magmatic heat source related to caldera-focused hydrothermal activity has been

imaged below Yellowstone Caldera at depths below 8 km (Miller and Smith, 1999;

Husen et ah, 2003). The only major hydrothermally active areas outside the caldera occur

in the Norris-Mammoth corridor. Notable features begin immediately north of the caldera

surrounding the Paintpot Hill, Gibbon Hill, and Gibbon River extracaldera rhyolites, and

extend northward through Norris Geyser Basin, Roaring Mountain (south of Obsidian

Cliff flow). Horseshoe Hill (north of Obsidian Cliff flow), and Mammoth Hot Springs

(Christiansen, 2001). In a study on hydrothermal fluid diseharges in the Norris-Mammoth

corridor, Kharaka et al. (2000) report chemical and isotopic values of thermal waters that

indicate separate magmatic sources for the Norris and Mammoth hydrothermal systems

based primarily on differences in magmatic ^He and CO 2 compositions. White et al.

(1988) suggest that the magmatic source for the entire corridor lies beneath Roaring

Mountain, which is at the center of the extracaldera dome field. Alternatively, Kharaka et

al. (2000) suggest that the heat and volatiles from the Mammoth system are derived from

a separate magma source of unknown location, emplaced near Mammoth at -400 ka

based on model U/Th ages of travertine deposits. The timing of emplacement of this

142

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. postulated Mammoth magma system is strikingly consistent with establishment of the

extracaldera magma system as early as 437 ka based on U/Pb zircon ages (Spell and

Nastanski, 2004).

Summary and Conclusions

The results presented herein support the hypothesis that the extracaldera Obsidian

Creek and Roaring Mountain member rhyolites were derived from a single long-lived and

coherent body of silicic magma that may have been established beneath the Norris-

Mammoth corridor as early as -450 ka based on zircon age distributions. Unrelated to

this magma system are the Cougar Creek and Riverside flows to the southwest, which

were originally assigned to the Roaring Mountain member. The magma reservoir from

which the remaining cogenetic rhyolites erupted was produced by accumulation of high-

silica rhyolite by effective crystal-melt segregation. This magma system persisted and

evolved for > 2 0 0 , 0 0 0 years, which suggests that it was thermally sustained by episodes

of mafic input throughout its eruptive period. Furthermore, Sr, Nd, and Pb isotopic data

have determined that the extracaldera magma system is unrelated to the adjacent sub-

caldera magma system that produced voluminous eruptions since 2.1 Ma, and thus

represents establishment of a new silicic magma system at Yellowstone.

The Obsidian Creek and Roaring Mountain member rhyolites record three significant

stages in the evolution of the magma system, (1) a 326-263 ka interval characterized by

initial eruption of porphyritic rhyolite followed by eruption of mingled lavas, which are

attributed to the interaction of intruding Swan Lake Flat basalt with a segregated melt

horizon, (2) a 263-134 ka interval characterized by a lull in mafic input, which results in

143

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cooling of the magma system, and continued evolution by fractional crystallization to

produce the remaining Obsidian Creek rhyolites, and (3) a rejuvenation event in which an

episode of focused mafic intrusion causes remelting of residual crystal mush and

generates a large-volume of less evolved magma that erupts over the 118-80 ka interval

as aphyric rhyolite and mingled lava. Pétrographie observations common to all mingled

lavas such as crenulated margins, exchange of phenocrysts across silicic-mafic interfaces,

and phenocryst reaction textures indicate that both silicic and mafic components were

liquid during interaction and eruption.

The temporally and spatially associated extracaldera Swan Lake Flat (SLF) basalts

likely played a significant role in the generation, evolution, and longevity of the Norris-

Mammoth magma system. Eruption of these basalts closely predate and are

contemporaneous with early erupted rhyolites and mingled lavas, which suggests that

they represent the composition of the intruding mafic magma during extracaldera

volcanism. Furthermore, geochemical and isotopic mixing models have shown that the

Swan Lake Flat basalts played a role in the production of the parental magma of the

extracaldera rhyolites, the hybrid dacite host of the Grizzly Lake mingled lava, and

hybrid basaltic-andesite enclaves from several mingled lavas. Evidence for rejuvenation

of the extracaldera magma system by renewed mafic input, and eruption of mingled lavas

during the earliest and most recent stages of volcanism indicate that the region beneath

the Norris-Mammoth corridor has been a focus of mafic intrusive activity throughout the

lifespan of this magma system.

Establishment of an evolving, long-lived silicic magma system north of Yellowstone

Caldera suggests that the extracaldera rhyolites may represent the onset of a new volcanic

144

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cycle. Whether or not this magma system has the potential to build to future explosive or

caldera-forming eruptions remains questionable and largely depends on the magnitude of

future mafic intrusive activity. Evidence for intruding mafic magma throughout the

duration of extracaldera volcanism strongly suggests that the Yellowstone melting

anomaly may have migrated north of Yellowstone Caldera beneath the Norris-Mammoth

corridor. This is supported by the fact that hydrothermal fluids in this area have high

^He/*He concentrations, which are thought to indicate the presence of a lower mantle

source. If the new Norris-Mammoth magma system does represent the most recent

migration of the Yellowstone melting anomaly, then it identifies the Norris-Mammoth

corridor as a highly potential site for future volcanism, including major explosive

eruptions. One the basis of the past three caldera-forming volcanic cycles at Yellowstone,

future eruptions are likely to be of a catastrophic nature.

145

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q. C g Q.

■D CD

C/) C/)

in Yellowstone C aldera 8

(O'

jQ Evolving Magma System Unrelated Rhyolites (0 n 2 3. 0. 3" CD % ■DCD m O 3 Q. C E a 3 3o u "O o\ o

CD Q. bmits

50 150 250 350 450 550 650 ■D Age (ka) CD

C/) C/) Figure 9.01. Cumulative probability diagram showing the eruptive periodicity of the Roaring Mountain and Obsidian Creek member rhyolites. The Cougar Creek (CC-7) and Riverside (RF-2) rhyolites are unrelated to the other extracaldera rhyolites, which represent an evolving magma system from 326 - 80 ka. Swan Lake Flat and Osprey basalts erupted in the same area from 350- 209 ka (Bennett, in progress) are shown with grey arrow. Timing of formation of Yellowstone Caldera (640 ka; Lanphere et al., 2002 ) is shown for reference. (a) (b)

Figure 9.02. Huppert and Sparks (1988) model for generation of a silicic magma system, (a) Inital stage where basalt penetrates the cold crust and may erupt to the surface as flows or form shields/cones, (b) Basalt induced heating initiates melting and generates magmas that can erupt, or form shallow intrusions, (c) Continued emplacement of basalt causes a large region of the crust to attain near melting temperatures and eventually large silicic magma bodies can form, ascend, and cause major caldera-forming eruptions. Basalts can erupt at the margins of the silicic focus of volcanism.

147

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (a) At crystal contents <45%, phenocrysts are kept in suspension by chaotic convection (low Reynolds number).

(b) At -45-50% crystals, the interstitial melt is rhyolitic and convection ceases. Melt extraction begins by a combination of hindered settling, micro-settling, and compaction of the mush.

^ V*Jif

(c) A rhyolitic horizon develops above the separated crystal mush and below an upper solidification front that results in crystallization along the cool roof.

Figure 9.03. Melt expulsion model for the generation of crystal-poor rhyolites from Bachmann and Bergantz (2004). (a) A large cooling, intermediate (andésite to dacite) magma body is capable of keeping crystals in suspension at crystallinities < 45%. (b) Enough crystallization occurrs (45-50%) to produce rhyolitic interstitial melt. Formation of a rigid crystal framework hinders convection and crystal-melt segregation begins, (c) A rhyolite horizon, from which crystal-poor rhyolites can erupt, develops above the residual crystal mush and below a crystallizing roof.

148

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. "DCD O Q. C g Q.

■D CD

C/) C/)

(a) 350 - 263 ka (b) 263- 134 ka (C) 1 3 4 -8 0 ka 8

(O'

' ^ . / V -V 3"3. CD w m ■DCD O caldera CQ. a 3o "O o

CD Q.

■D LEGEND CD m a Roaring Mtn. rhyolites Obsidian Creek rhyolites Swan Lake Flat basalt Crystal mush Rhyolitic magma Basalt Intrusions (/) Figure 9.04. Schematic diagram depicting evolution of the extracaldera rhyolite magma system. See text for detailed explanations.

(a) Stage 1 depicts establishment of the (b) Stage 2 is characterized by a lull in (c) Stage 3 shows rejuvenation of the magma system showing early SLF basalt mafic input which results in cooling and magma system due to renewed mafic input eruptions (yellow) followed by eruption of chemical evolution by fractional which causes remelting of residual crystal rhyolite and mingled lavas that periodically crystallization to produce the remaining mush and results in eruption of voluminous tap a segregated melt horizon. Obsidian Creek member rhyolites. aphyric magma. APPENDIX A

^®Ar/^^Ar DATA TREATMENT AND ANALYTICAL DATA

150

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Dating of Young Samples

Since the ''^Ar/^^Ar method of dating is based on the accumulation of radiogenic

argon ('*°Ar*) from the decay of potassium ('*®K) over time, a major limitation to the

technique (as well as to all isotopic dating systems) is that age determinations decrease in

precision as samples approach younger timescales. With young samples the measured

''^Ar* can be low enough that the corrections and calibrations applied during data

reduction cause large uncertainties in the ealculated ages. This was regarded as a

potential problem since the previously reported K/Ar ages (399 - 80 ka; Obradovich,

1992) from the extraealdera rhyolites approach the limits of the "^°Ar/^^Ar teehnique.

Young samples therefore demand maximizing the sample gas yields while minimizing

blank corrections.

Blanks are routinely measured during laser fusion analyses to monitor and correet for

background levels of each measured isotope of Ar in the extraetion line. For young

samples age calculations are sensitive to blank corrections even though measured blanks

were remarkably low. One approach to lessen the effeets of blank values on a particular

sample was to average the blank measured during laser fusion analyses over a period of

time. Another approach is to increase the amount of gas released by fusing multiple

sanidine crystals/ glass fragments versus single crystals/fragments. Multiple

crystal/fragment fusion would increase sample gas yields and help determine reliable

ages. It should be noted that when multiple sanidine erystals are analyzed it deereases the

ability to clearly differentiate between xenoerystic and juvenile material. In the case that

multiple crystal/fragment analyses eontinued to output a low gas yield, furnace step

heating of bulk sanidine/glass separates was be attempted.

151

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A second limitation to dating the young rhyolites involved the precise dating of

aphyrie lavas in which glass was the only material suitable for analysis. Sanidine was the

most suitable mineral for dating the porphyritie rhyolites because it has low amounts of

initial Ar (important for young samples) and its lattice sites are eapable of holding

measurable amounts (up to 12 - 14%) of potassium. Voleanie glass commonly yields

anomalous ages beeause it laeks an internal atomic lattice structure, which can lead to an

uptake of atmospheric argon and a loss of potassium (through K-Na exchange) and

radiogenic argon during hydration/devitrification events. In order to limit this potential

problem, apparently fresh glasses from one porphyritic rhyolite vitrophyre (CC-7), and

from two aphyrie rhyolites (OC-5 and CS-1) were elosely examined for signs of

hydration and/or devitrification. Glass from CC-7 was chosen for dating beeause the

porphyritic sample from this dome was altered and sanidine was partly replaeed by

secondary calcite. Loss on ignition (LOI) measurements listed in Table 1 of Appendix B

were eonducted to determine the amount of hydration in glasses. The total loss on

ignition measurements were highest for glasses, but did not exceed ~ 0 . 6 %, indicating

that the glasses to be analyzed were only slightly hydrated. Pétrographie results (Chapter

5) showed sparse spherulites in the CC-7 vitrophyre indieating slight devitrifieation, but

no devitrification was evident in the OC-5 and CS-1 rhyolites. Laser fusion analysis of

glass was attempted in addition to incremental step heating to eompare each method’s

ability to determine reliable ages.

152

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Data Treatment

Statistical methods from ISOPLOT (Ludwig, 2001), an plug-in for Microsoft Excel,

were used to evaluate and display the measured sample ages and standard deviations

from the '''’Ar/^^Ar analyses. All ages reported in data tables from Appendix A were

ealeulated using the following statistical methods. Homogeneity of analyzed erystal

populations from laser fusion analysis was assessed by plotting data on eumulative

probability curves that allow visual identifieation of mixed age populations of juvenile

phenocrysts and outlying xenocrysts. Any analyses greater than 2a from the sample

mean of the entire population were omitted and a new refined mean was ealeulated. An

error-weighted average of the refined dataset is ealeulated and displayed using the

Weighted Average extension of ISOPLOT. A preferred weighted mean age was taken

when a refined population was just below the critical MSWD value (Wendt and Carl,

1991).

Isochron ages from laser fusion analyses were derived from an inverse isochron

diagram, whieh plots data as ^Âr/'Âr versus ^Âr/Âr on a best-fit line. A routine of

eliminating analyses that contribute to a MSWD greater than the critical value for that

population was repeated until an aeceptable MSWD was achieved (Wendt and Carl,

1991). Meaningful isoehrons are defined by four or more data points. The isochron

method is useful because it extracts meaningful information regarding the sample age at

the x-intereept (^^Ar/°Ar) and the composition of trapped argon at the y-intercept

(^^Ar/^Ar). For laser fusion analyses, an isochron age is preferred if obtained, and a

weighted mean age is considered the next most reliable.

153

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Data from a furnace step heat analysis are displayed on age spectrum and isochron

diagrams. Criteria for a meaningful furnace step heat age are a well-defined plateau

and/or a well-defined isochron. A plateau is defined when gas fractions from contiguous

steps yield indistinguishable ages and total > 50% of the ^^Ar released. The plateau age is

ealeulated as the error-weighted mean of the steps that define the plateau (McDougall and

Harrison, 1999). Isoehrons and isochron ages from furnace step heat analyses are

determined as explained previously, however data defining a valid isoehron must

represent contiguous step analyses. For step heating analyses, an isochron age is

eonsidered the most reliable if obtained. If no isochron is obtained then a plateau age is

preferred and a total gas age is eonsidered the least reliable.

Comparison of'^‘^Ar/^^Ar Ages to K/Ar Ages

Geoehronology results (Table 6.01) indieate that several '’^Ar/^^Ar dates obtained in

this study are not eonsistent (at 2o) with the previous K/Ar dates from Obradovich

(1992). The K/Ar age (90 ± 4) of sanidine from the Gibbon River rhyolite is slightly

younger than the preferred '*'^Ar/^^Ar age (118 + 20) of sanidine from sample GR-2. The

younger K/Ar age is interpreted to reflect incomplete degassing of radiogenie argon in

sanidine during K/Ar analysis. K/Ar ages of glass from the Cougar Creek (399 ± 6 ) and

Obsidian Cliff (183 ± 6 ) rhyolites are significantly older than the preferred '^'^Ar/^^Ar ages

of CC-7 (368 ± 4, sanidine) and OC-5 (106 ± 2, glass), respeetively. The older K/Ar ages

from the Cougar Creek and Obsidian Cliff rhyolites are likely due to open system

behavior of glass. It is unknown if the K/Ar age of glass from the Crystal Spring rhyolite

is well eonstrained, however it is used in this study sinee the "^^Ar/^^Ar step heating

154

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. analysis of glass did not produce a reliable age (Appendix A). K/Ar ages of sanidine from

the Gibbon Hill and Willow Park rhyolites agree with the preferred '''^Ar/^^Ar ages of

sanidine from GH-2, WP-1. ''^Ar/^^Ar ages were obtained from all other rhyolites that

were previously undated.

155

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q. C 8 Q.

■D CD

C/) C/)

CD CS-1, glass, 42.33 mg, J = 0.0002327 ± 0.32% 4 amu discrimination - 1.01806 ± 0.30%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.59%, 39/37Ca = 0.000783 ± 8.75% 3. 3" CD step T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* % 39Ar risd Ca/K 40Ar*/39ArK Age (ka) is.d. 1 675 12 1,539 1.679 1.459 58.881 441.465 -1.0 10.1 0.817712094 -0.0762 -32.0 -10.0 CD ■D 2 750 12 0.666 1.197 0.758 41.935 183.846 -4.8 7.2 0.818544913 -2.0957 -880.0 -6.0 O Q. 3 810 12 0.540 1.044 0.682 37.272 151.48 -3.1 6.4 0.803231648 -0.1239 -52.0 -7.0 C a 4 870 12 0.540 1.040 0.660 36.000 147.447 -5.9 6.2 0.828432524 -0.2382 -100.0 -7.0 O 5 930 12 0.565 1.085 0.720 38.507 156.536 -4.4 6.6 0.808004328 -0.1763 -74.0 -7.0 3 ■D 6 990 12 0.574 1.144 0.717 39.640 162.818 -1.9 6.8 0.82759636 -0.0786 -33.0 -11.0 O 7 1045 12 1.213 2.692 1.519 82.949 345.035 -1.8 14.2 0.930686809 -0.0738 -31.0 -7.0 8 1100 12 1.413 3.401 1.793 99.381 406.869 -0.5 17.0 0.981407906 -0.0214 -9.0 -7.0

CD 9 1150 12 1.613 3.754 1.854 103.536 466.398 -0.1 17.7 1.039816721 -0.0048 -2.0 -7.0 Q. 10 1200 12 0.376 2.367 0.440 24.856 114.07 5.2 4.3 2.732385252 0.2358 99.0 10.0 11 1250 12 0.385 5.388 0.369 20.907 118.461 4.8 3.6 7.404918901 0.2644 111.0 9.0 Cumulative %39Ar risd = 100.0 Total gas age = -81.1 4.6 released, error in age includes 0.5% J error, all errors 1 sigma No Plateau ■D Note: isotope beams in mV, risd = CD (36Ar through 40Ar are measured beam intensities, corrected for decay for the age calculations) No Isochron

C/) C/) ■o I I

(g(/) 3o'

CD 8 5 c5'

3 CD OC-5, 002 laser fusion glass, J = 0.00023760 ± 0.31% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75% C p. 3" CD Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 6 0.557 0.321 21.459 11.136 45.4 0 1600 0.026 0.1112248 193.2722 81.0 8.0 ■o 2 1600 6 0.033 1.165 0.715 46.943 17.942 56.8 0.10634338 195.7130 82.0 1.0 3 1600 6 0.024 1.329 0.807 53.925 18.781 75.1 0.10560637 237.4129 99.0 1.0 c1 a 4 1600 6 0.024 0.656 0.368 25.454 10.779 50.2 0.11043421 171.3665 72.0 3.0 o 5 1600 6 0.058 1.035 0.620 42.146 24.626 37.2 0.10522992 203.0432 85.0 7.0 3 L/l ■o -J 6 1600 6 0.030 0.807 0.496 31.881 13.787 48.3 0.1084669 178.6563 75.0 3.0 o 7 1600 6 0.017 0.723 0.415 28.169 10.802 74.3 0.10998226 232.4866 97.0 2.0 8 1600 6 0.031 0.894 0.533 36.061 17.110 58.0 0.10623194 247.2818 103.0 2.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 86.8 10.8 & (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 96.0 6.7 (omit# 1,2,3,8) co Wtd mean = 99.1 0.8 (omit #1,2,3,8) % Isochron = 81.0 17.5 (gC/) o' 3 ■o I I

(g(/) 3o'

8 c5'

3 OC-5, glass, 38.10 mg, J = 0.0002376 ± 0.31% CP 4 amu discrimination = 1.01806 ± 0.30%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 + 17.59%, 39/37Ca = 0.000783 ± 8.75%

step T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* % 39Ar risd Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 675 12 0.393 0.759 0.566 33.093 122.74 8.3 6.6 0.096902425 316.9808 131.0 6.0 2 750 12 0.064 0.834 0.564 37.299 25.703 35.3 7.4 0.094448136 234.9490 98.0 1.0 ■oCP 3 800 12 0.032 0.957 0.630 42.693 16.114 57.0 8.5 0.094707416 198.1550 83.0 1.0 4 820 12 0.020 0.867 0.555 38.101 13.642 77.1 7.6 0.096141663 249.7525 104.0 2.0 Ic 5 840 12 0.022 0.788 0.539 35.403 12.31 68.6 7.1 0.094040469 210.3857 88.0 2.0 a 6 865 12 0.018 0.794 0.529 35.714 12.14 79.2 7.1 0.093931364 237.4129 99.0 2.0 o Ul 3 00 7 890 12 0.014 0.801 0.528 36.312 12.226 73.5 7.2 0.093198918 210.3857 88.0 2.0 ■o 8 950 12 0.020 1.114 0.775 15.066 16.335 81.2 3.0 0.312424002 844.7691 330.0 5.0 o 9 1000 12 0.022 1.176 0.801 54.386 17.065 78.3 10.8 0.0913584 222.6504 93.0 1.0 10 1050 12 0.020 1.131 0.760 51.166 15.858 80.6 10.2 0.093392002 222.6504 93.0 3.0 & 11 1100 12 0.014 0.907 0.618 41.775 13.829 91.7 8.3 0.091731641 267.0856 111.0 1.0 , 12 1150 12 0.021 0.860 0.581 38.831 13.894 73.6 7.7 0.093572547 222.6504 93.0 2.0 13 1220 12 0.011 0.629 0.432 28.782 11.55 98.2 5.7 0.092333243 324.5129 134.0 3.0 o c 14 1400 12 0.032 0.303 0.187 12.708 13.216 20.7 2.5 0.100738373 132.6919 56.0 3.0 Cumulative %39Ar risd = 100.0 Total gas age : 105.6 1.0 Note: isotope beams in mV, risd = released, error In age Includes 0.5% J error, all errors 1 sigma No plateau (36Ar through 40Ar are measured beam intensities, corrected for decay for the age calculations) No isochron (/)w 3o' CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci' GR-2, C02 laser fusion sanidine, J = 0.00023592 ± 0.29% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) is.d . 3 1 1600 4 0.132 1.154 0.932 72.061 71.431 48.2 0.06673443 496.9417 200.0 4.0 3" CD 2 1600 4 0.092 1.354 1.053 83.172 56.253 55.2 0.06784003 380.2786 155.0 4.0 4 0.105 0.977 0.793 61.565 CD 3 1600 65.465 38.3 0.06613099 418.8434 170.0 3.0 ■D 4 1600 4 0.143 0.884 0.682 52.739 O 61.545 34.3 0.06984979 408.5281 166.0 3.0 Q. 5 1600 4 0.078 0.862 0.684 52.844 39.709 46.2 0.06797608 347.1142 142.0 2.0 C a 6 1600 4 0.066 1.057 0.846 65.802 40.609 56.5 0.06693917 349.6568 143.0 2.0 3o 7 1600 4 0.076 1.175 0.784 60.018 45.034 54.3 0.08158352 413.6829 168.0 2.0 "O 8 1600 4 0.127 0.834 0.708 53.989 50.105 28.3 0.06437317 263.9930 109.0 2.0 O 9 1600 4 0.048 1.053 0.834 68.839 33.961 63.7 0.06374379 311.6649 128.0 3.0 10 1600 4 0.133 1.039 0.745 57.257 51.696 27.2 0.07561925 246.5559 102.0 2.0 CD 11 1600 4 0.104 0.700 0.556 43.518 45.194 28.1 0.06703067 296.5563 122.0 7.0 Q. 12 1600 4 0.114 1.165 0.843 67.178 54.582 35.5 0.07226764 294.0431 121.0 4.0 Note: isotope beams In mV risd = released, error In age Includes J error, all errors 1 sigma Mean + s.d. = 143.8 27.6 (36Ar through 40Ar are measured beam Intensities, corrected for decay in age calculations) Mean ± s.d. = 142.5 0.5 ■D (omit# 1,2,3,4,7,8,9,10,11,12) CD Wtd mean = 142.5 1.4

C/) (omit# 1,2,3,4,7,8,9,10,11,12) C/) Isochron = 118.0 10.0 CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci'

GR-2, sanidine, 36,32 mg, J = 0.0002359 + 0,29% 4 amu discrimination = 1.01806 ± 0.30%, 40/39K = 0.0001 ± 100.0%, 36/37Ca : 0.000397 ± 17.59%, 39/37Ca = 0.000783 ± 8.75%

3 step T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* % 39Ar risd Ca/K 40Ar*/39ArK Age (ka) Is.d. 3" CD 1 750 12 0.245 0.183 0.138 6.948 72.167 3.3 0.9 0.110140525 349.6865 143.0 54.0 2 810 12 0.160 0.233 0.166 10.832 47.195 2.8 1.4 0.089949861 119.1294 50.0 7.0 ■DCD O 3 870 12 0.072 0.352 0.238 17.789 25.089 20.4 2.3 0.082745315 276.5130 114.0 5.0 Q. 4 930 12 0.083 0.490 0.366 27.834 29.088 20.5 3.6 0.073615887 206.9702 86.0 5.0 C a 5 990 12 0.080 0.648 0.508 39.706 32.162 31.5 5.1 0.068244752 249.0639 103.0 3.0 3o 6 1045 12 0.081 0.739 0.657 49.673 32.194 30.7 6.4 0.062211926 192.2081 80.0 3.0 "O o 7 1100 12 0.078 0.838 0.734 58.741 35.238 39.7 7.6 0.059655701 231.6831 96.0 1.0 O 8 1150 12 0.067 0.901 0.805 63.347 31.874 44.4 8.2 0.059476853 211.9019 88.0 1.0 9 1200 12 0.096 1.055 0.985 76.297 44.886 41.7 9.8 0.057822122 239.1238 99.0 1.0 CD 10 1250 12 0.224 2.043 1.883 149.684 128.137 50.9 19.3 0.057074474 452.5650 183.0 1.0 Q. 11 1300 12 0.279 2.343 2.280 177.374 152.647 48.3 22.8 0.055237119 431.8064 175.0 1.0 12 1350 12 0.082 0.964 0.949 74.916 47.258 50.7 9.6 0.053808509 296.5815 122.0 3.0 13 1400 12 0.059 0.313 0.295 23.259 27.297 35.8 3.0 0.056273266 357.3235 146.0 5.0 ■D Cumulative %39Ar risd = 100.0 Total gas age = 133.6 1.0 CD Note: isotope beams in mV, risd = released, error in age includes 0.5% J error, all errors 1 sigma No plateau (36Ar through 40Ar are measured beam intensities, corrected for decay for the age calculations) No isochron C/) C/) CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci'

CC-7, C02 laser fusion multiple glass fragments, J = 0.00023706 ± 0.28% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca ■: 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

3 Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 3" 1 1600 6 0.029 0.203 0.091 5.741 12.703 41.2 0.15488055 813.1147 318.0 15.0 CD 2 1600 6 0.034 0.369 0.155 10.861 20.149 57.8 0.14881417 1064.6253 406.0 12.0 ■DCD 3 1600 6 0.020 0.234 0.108 6.538 10.760 58.8 0.15676877 829.8761 324.0 13.0 O 4 1600 6 0.025 0.247 0.106 7.449 13.012 53.9 0.14524002 846.6934 330.0 25.0 Q. C 5 1600 6 0.035 0.480 0.204 14.751 23.759 63.7 0.14253017 1035.4228 396.0 9.0 a 6 1600 6 0.108 0.330 0.158 8.809 39.500 21.9 0.16408797 1035.4228 396.0 9.0 3o Ch "O 7 1600 6 0.038 0.447 0.211 13.752 24.368 60.7 0.14237334 1096.9355 417.0 16.0 O 8 1600 6 0.016 0.142 0.062 3.804 7.879 58.0 0.16350732 934.4795 361.0 23.0 9 1600 6 0.042 0.475 0.207 14.185 24.641 56.0 0.14667353 980.3821 377.0 12.0

CD 10 1600 6 0.033 0.415 0.186 12.110 19.621 58.3 0.15010385 920.2182 356.0 14.0 Q. 11 1600 6 0.075 0.616 0.292 18.953 39.584 48.0 0.14236058 1058.7718 404.0 5.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 371.4 34.0 (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 393.9 17.6 ■D (omit# 1,3,4,10) CD Wtd mean = 399.3 3.6 (omit# 1,3,4,10) C/) C/) Isoctiron = 391.0 14.0 CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci'

CC-7, glass, 61.52 mg, J = 0.0002371 ± 0.28% 4 amu discrimination = 1.01806 ± 0.30%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.59%, 39/37Ca = 0.000783 ± 8.75%

step T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* % 39Ar risd Ca/K 40Ar*/39ArK Age (Ma) Is.d. 3 3" 1 675 12 1.877 2.100 1.514 78.726 626.777 13.3 8.5 0.113553444 1200.8587 452.0 10.0 CD 2 750 12 0.228 2.222 1.259 83.334 147.636 56.5 9.0 0.113506576 1120.3689 425.0 3.0 ■DCD 3 800 12 0.082 2.451 1.380 92.901 112.853 81.1 10.1 0.112310914 1093.8055 416.0 3.0 O 4 820 12 0.045 2.255 1.233 84.043 92.715 88.7 9.1 0.114220561 1082.0421 412.0 2.0 Q. C 5 840 12 0.037 2.104 1.168 80.430 86.49 90.6 8.7 0.111359323 1073.2366 409.0 2.0 a 6 865 12 0.031 2.090 1.178 79.160 84.487 92.5 8.6 0.112393075 1087.9205 414.0 2.0 3o 7 890 12 0.032 2.043 1.115 76.187 81.376 91.8 8.3 0.114152854 1079.1053 411.0 4.0 "O S 8 950 12 0.041 2.623 1.454 98.780 105.201 91.3 10.7 0.113039025 1076.1701 410.0 2.0 O 9 1000 12 0.042 2.654 1.362 93.542 99.997 90.5 10.2 0.12077983 1070.3046 408.0 3.0 10 1050 12 0.036 2.329 1.079 72.647 80.187 90.3 7.9 0.136475282 1099.6971 418.0 3.0 CD 11 12 0.659 45.035 7.0 Q. 1100 0.033 1.569 52.029 86.2 4.9 0.148312078 1082.0421 412.0 12 1150 12 0.024 0.712 0.331 22.393 29.69 83.9 2.4 0.135353585 1105.5951 420.0 5.0 13 1220 12 0.030 0.333 0.156 10.131 17.868 57.8 1.1 0.139924759 905.8436 351.0 4.0 14 1400 12 0.038 0.167 0.062 4.067 18.614 50.2 0.4 0.17480325 1657.3597 599.0 19.0 ■D Cumulative %39Ar risd = 100.0 Total gas age = 417.0 2.5 CD Note: isotope beams in mV, risd = released, error in age includes 0.5% J error, all errors 1 sigma Plateau age = 411.4 1.9 (36Ar through 40Ar are measured beam intensities, corrected for decay for the age calculations) (steps 3-9) C/) C/) Isochron age = 378.0 2.0 (steps 7-12) CD ■ D O Q. C g Q.

■D CD

C/) C/)

(O'

CC-7s, sanidine,13.82 mg, J = 0.000237 ± 0.28% 4 amu discrimination = 1.01700 ± 0.03%, 4Q/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

step T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* % 39Ar risd Ca/K 40Ar*/39ArK Age (ka) Is.d. 3. 1 625 12 0.206 0.161 0.068 2.155 69.08 14.1 0.6 0.24767258 7657.2282 23.0 3" 1867.0 CD 2 700 12 0.053 0.253 0.062 3.090 17.57 15.4 0.9 0.271434141 810.5316 317.0 68.0 3 775 12 0.047 0.469 0.093 6.262 18.69 31.7 1.8 0.248289976 875.0684 340.0 20.0 ■DCD O 4 850 12 0.060 0.810 0.159 11.778 26.53 38.6 3.4 0.227987073 849.7168 331.0 27.0 Q. 5 920 12 0.057 1.268 0.259 20.216 33.90 55.6 5.9 0.207930481 929.0054 359.0 3.0 C a 6 990 12 0.064 1.752 0.419 31.537 43.97 61.8 9.2 0.184164006 877.8930 341.0 2.0 3o 7 1055 12 0.068 2.028 0.521 40.945 51.20 65.3 11.9 0.164193324 838.4900 327.0 3.0 "O 8 1110 12 0.061 1.956 0.576 44.762 52.78 70.5 13.0 0.144858926 855.3396 333.0 5.0 o 9 1155 12 0.044 1.595 0.538 42.885 48.20 77.9 12.5 0.12329298 877.8930 341.0 4.0 10 1200 12 0.042 1.304 0.521 40.324 46.29 78.2 11.8 0.107200043 897.7095 348.0 2.0 CD 11 1245 12 0.151 1.143 0.617 46.234 87.799 51.0 13.5 0.081952544 1032.7729 395.0 4.0 Q. 12 1290 12 0.300 0.936 0.441 34.935 41.260 51.0 10.2 0.088816504 995.0622 382.0 3.0 13 1340 12 0.019 0.510 0.180 13.529 21.623 98.2 3.9 0.124964822 977.7487 376.0 3.0 14 1400 12 0.026 0.270 0.058 4.431 19.585 82.8 1.3 0.202001863 2413.2375 816.0 17.0 ■D Cumulative %39Ar risd = 100.0 Total gas age = 368.2 2.0 CD Note: isotope beams In mV, rIsd = released, error in age includes 0.5% J error, all errors 1 sigma No Plateau (36Ar through 40Ar are measured beam Intensities, corrected for decay for the age calculations) Isochron Age = 286.0 28.0 C/) C/) CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci' RF-2, C02 laser fusion single crystal sanidine, J = 0.00023620 ± 0.32% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) Is.d. 1 1600 6 0.088 0.687 0.632 48.189 88.174 73.4 0.06251979 1540.9886 560.0 3.0 3 2 1600 6 0.080 0.978 1.020 80.629 123.243 83.1 0.0531931 1455.2075 533.0 6.0 3" CD 3 1600 6 0.074 0.537 0.571 44.405 78.928 75.5 0.05303347 1531.3938 557.0 3.0 4 1600 6 0.068 0.753 0.487 37.370 68.601 74.3 0.08836565 1553.8064 564.0 3.0 CD ■ D 5 1600 6 0.047 0.454 0.479 36.488 58.445 80.5 0.05456493 1445.7553 530.0 4.0 O Q. 6 1600 6 0.040 0.556 0.435 33.844 55.420 83.2 0.07204484 1537.7886 559.0 3.0 C a 7 1600 6 0.041 0.307 0.388 29.626 47.502 78.4 0.04544354 1398.7292 515.0 4.0 3O 8 1600 6 0.020 0.472 0.464 37.935 52.794 92.7 0.05456444 1448.9043 531.0 5.0 "O S 9 1600 6 0.039 0.386 0.421 33.335 52.378 81.7 0.05078018 1439.4625 528.0 4.0 O 10 1600 6 0.050 0.492 0.530 42.032 65.461 80.4 0.05133251 1411.2313 519.0 4.0 11 1600 6 0.038 0.515 0.548 43.552 66.007 86.1 0.05185691 1480.4904 541.0 3.0 CD 12 1600 6 0.022 0.289 0.245 18.596 29.101 84.1 0.06815342 1426.8979 524.0 14.0 Q. 13 1600 6 0.051 0.498 0.407 30.162 51.659 74.4 0.07240674 1426.8979 524.0 4.0 14 1600 6 0.125 0.479 0.574 43.866 91.575 61.8 0.04788666 1477.3239 540.0 3.0 Note; isotope beams in mV risd = released, error in age inciudes J error, all errors 1 sigma Mean ± s.d. = 537.5 15.8 ■D (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 535.0 4.6 CD (omit# 1,3,4,6,7,9,10,12,13) Wtd mean = 537.5 2.2 C/) C/) (omit# 1,3,4,6,7,9,10,12,13) Isochron = 525.8 3.3 CD ■ D O Q. C g Q.

■D CD

C/) C/)

CD 8

GH-2, C 02 laser fusion single crystal sanidine, J = 0.00023498 ± 0.28% 4 amu discrimination = 1.01951 ± 0.13%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca ; : 0.000783 ± 8.75%

CD Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 1600 6 0.060 0.376 0.429 33.404 44.475 66.5 0.03846539 934.1141 358.0 3.0 3. 2 1600 6 0.070 0.368 0.400 29.796 29.780 38.1 0.04220569 368.9649 150.0 3.0 3" CD 3 1600 6 0.309 0.361 0.432 30.083 98.076 9.9 0.04100786 330.6727 135.0 7.0 4 1600 6 0.074 0.260 0.283 21.156 27.450 27.6 0.04199726 343.4014 140.0 4.0 ■DCD O 5 1600 6 0.057 0.356 0.409 29.684 24.817 37.3 0.04098346 338.3057 138.0 4.0 Q. 6 1600 6 0.056 0.338 0.327 23.860 22.999 41.1 0.04840922 325.5911 133.0 4.0 C a 7 1600 6 0.097 0.267 0.292 21.640 35.275 24.6 0.04216335 394.6705 160.0 3.0 O 3 LAON 8 1600 6 0.185 0.286 0.328 23.324 60.739 13.7 0.0419029 353.6098 144.0 7.0 "O 9 1600 6 0.266 0.514 0.592 43.573 89.584 15.1 0.04031128 315.4447 129.0 3.0 O 10 1600 6 0.062 0.384 0.432 32.326 27.110 39.5 0.04059385 315.4447 129.0 4.0 11 1600 6 0.125 0.466 0.485 36.939 47.34 26.2 0.0431104 333.2156 136.0 3.0 CD Q. 12 1600 6 0.051 0.259 0.275 20.487 19.675 31.9 0.04320188 280.1091 115.0 6.0 13 1600 6 0.043 0.372 0.393 30.234 50.107 80.7 0.04204637 1497.7364 544.0 5.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 185.5 119.4 (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 135.5 4.9 ■D (omit #1,2,7,12,13) CD Wtd mean = 134.3 1.4

C/) (omit #1,2,7,12,13) C/) Isochron = 134.3 2.6 "OCD O Q. C g Q.

"O CD

C/) C/)

PH-5, C02 laser fusion single crystal sanidine, J = 0.00023648 ± 0.30% 4 amu discrimination = 1.01951 ± 0.13%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

CD Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 1600 6 0.052 0.424 0.438 33.093 32.366 62.8 0.0452396 617.7769 246.0 6.0 3. 2 1600 6 0.054 0.326 0.370 27.907 29.529 56.8 0.04124706 599.0086 239.0 7.0 3" 3 1600 6 0.017 0.323 0.326 24.813 18.610 93.2 0.04596342 669.0882 265.0 7.0 CD 4 1600 6 0.039 0.428 0.448 36.265 30.773 67.8 0.04167204 574.9847 230.0 3.0 "OCD 5 1600 6 0.118 0.744 0.672 51.140 59.608 44.2 0.0513691 529.9318 213.0 2.0 O 6 1600 6 0.023 0.294 0.279 22.665 20.128 74.4 0.04580161 636.6182 253.0 6.0 Q. C 7 1600 6 0.050 0.417 0.472 35.562 31.367 57.6 0.04140364 503.6277 203.0 3.0 a O 8 1600 6 0.036 0.453 0.457 36.238 29.085 69.0 0.04413904 548.4316 220.0 3.0 3 Os "O G\ 9 1600 6 0.083 0.493 0.489 37.831 43.026 46.3 0.04601381 535.2101 215.0 2.0 O 10 1600 6 0.300 0.468 0.475 37.951 27.680 74.2 0.04354231 532.5702 214.0 2.0 11 1600 6 0.183 0.571 0.565 44.082 35.108 33.7 0.0457366 596.3334 238.0 2.0

CD 12 1600 6 0.045 0.557 0.565 44.082 35.108 66.8 0.0446152 532.5702 214.0 1.0 Q. 13 1600 6 0.141 0.517 0.516 38.723 61.33 34.5 0.0471423 564.3458 226.0 2.0 14 1600 6 0.040 0.371 0.377 29.430 27.714 62.8 0.04451154 585.6472 234.0 4.0 15 1600 6 0.031 0.460 0.458 36.162 26.731 72.0 0.04491531 522.0252 210.0 2.0 "O Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 228.0 17.7 CD (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 214.3 3.0 (omit# 1,2,3,4,6,7,11,13,14) C/) C/) Wtd mean = 213.9 0.9 (omit# 1,2,3,4,6,7,11,13,14) Isochron = 208.1 4.9 CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci' LD-1, C02 laser fusion single crystal sanidine, J = 0.00023728 ± 0.28% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca : 0.000783 ± 8.75%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) Is.d. 1 1600 6 0.070 0.526 0.523 42.124 39.786 54.5 0.04621833 509.7601 208.0 2.0 2 1600 6 0.040 0.364 0.372 28.328 25.325 63.6 0.04756021 546.2281 222.0 2.0 3 3" 3 1600 6 0.044 0.471 0.513 39.345 33.046 69.0 0.04430873 572.4505 232.0 3.0 CD 4 1600 6 0.050 0.447 0.468 35.672 32.707 62.9 0.0463808 569.8217 231.0 2.0 ■DCD 5 1600 6 0.030 0.283 0.324 25.236 20.943 70.6 0.04150721 551.4610 224.0 8.0 O 6 1600 6 0.074 0.315 0.334 25.628 32.502 39.5 0.04549398 489.0479 200.0 4.0 Q. C 7 1600 6 0.031 0.346 0.377 20.818 23.414 72.7 0.06151725 525.3546 214.0 7.0 a O 8 1600 6 0.039 0.410 0.326 25.174 23.181 61.3 0.06028251 538.3897 219.0 4.0 3 Cn 9 1600 6 0.029 0.290 0.241 18.615 15.706 61.3 0.05766262 463.2866 190.0 8.0 "O O 10 1600 6 0.054 0.607 0.583 46.449 41.363 68.2 0.0483694 612.0576 247.0 5.0 11 1600 6 0.049 0.546 0.572 43.214 35.641 67.1 0.04676559 548.8438 223.0 2.0 12 1600 6 0.085 0.627 0.729 56.524 53.487 58.2 0.04105747 556.6996 226.0 5.0 CD Q. 13 1600 6 0.049 0.556 0.606 46.934 37.052 68.5 0.04384753 535.7798 218.0 5.0 14 1600 6 0.104 0.463 0.460 35.104 47.450 40.3 0.04881832 548.8438 223.0 2.0 15 1600 6 0.034 0.357 0.401 31.638 25.049 70.8 0.04176542 538.3897 219.0 3.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 219.7 13.0 ■D CD (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 220.9 3.5 (omit# 1,3,4,6,9,10) C/) C/) Wtd mean = 221.9 1.1 (omit# 1,3,4,6,9,10) Isochron = 226.0 5.5 CD ■ D O Q. C g Q.

■D CD

C/) C/)

CD 8

GLM-2, C02 laser fusion multiple crystal sanidines, J = 0.00023357 ± 0.54% 4 amu discrimination = 1.01700 ± 0.03%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75% 3. 3" CD Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 1600 6 0.336 0.909 0.692 50.100 130.560 25.1 0.07203826 696.7022 272.0 5.0 ■DCD O 2 1600 6 0.039 0.572 0.237 18.160 22.476 52.0 0.12506135 614.6063 242.0 8.0 Q. 3 1600 2 0.019 0.489 0.291 22.856 22.752 85.6 0.08494681 902.2654 345.0 4.0 C a 4 1600 6 0.137 1.282 0.476 36.672 65.373 42.5 0.13880269 794.2222 307.0 4.0 O Ch 5 1600 6 0.058 0.547 0.320 25.930 32.357 55.3 0.08375739 688.4310 269.0 19.0 3 0 0 "O 6 1600 6 0.080 0.799 0.538 42.370 49.290 57.9 0.07487307 691.1866 270.0 3.0 O 7 1600 6 0.037 0.542 0.404 32.167 29.526 73.3 0.06689986 663.6999 260.0 4.0 8 1600 6 0.074 0.514 0.353 27.405 36.779 49.2 0.07446816 663.6999 260.0 11.0 CD Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 278.1 30.5 Q. (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 266.2 5.2 (omit #2, 3, 4) Wtd mean = 267.4 2.4 "O isochron = 263.3 3.4 CD

C/) C/) CD ■ D O Q. C 8 Q.

■D CD

C/) C/)

8 ci' GRM-1, C02 laser fusion multiple crystal sanidines, J = 0.00023631 ± 0.31%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 3 1 1600 6 0.341 2.991 1.134 85.802 157.767 39.0 0.13019538 773.9025 303.0 4.0 3" 2 1600 6 1.763 1.885 1.487 91.848 577.480 11.3 0.07665002 773.9025 303.0 2.0 CD 3 1600 6 0.071 2.252 0.953 74.287 76.024 74.6 0.11322185 815.6953 318.0 2.0 ■DCD 4 1600 6 0.142 1.632 0.889 67.569 85.216 52.4 0.09020794 699.5455 276.0 2.0 O 5 1600 6 0.121 2.512 0.929 71.621 87.452 61.0 0.13099536 796.1487 311.0 2.0 Q. C 6 1600 6 0.079 0.738 0.497 39.114 49.549 57.3 0.07046834 754.5178 296.0 4.0 a 7 1600 6 0.086 1.058 0.513 39.808 52.740 56.1 0.09926324 776.6779 304.0 9.0 3O Os "O 8 1600 6 0.308 0.415 0.335 24.081 106.225 16.5 0.06436397 779.4548 305.0 13.0 O 9 1600 6 0.093 0.743 0.585 43.875 57.752 56.3 0.06324713 776.6779 304.0 8.0 10 1600 6 0.037 0.807 0.601 48.917 44.415 81.1 0.06161446 760.0486 298.0 3.0 0.410 0.369 25.009 90.777 18.0 0.06122889 691.3521 273.0 13.0 CD 11 1600 6 0.259 Q. 12 1600 6 0.195 1.052 0.678 50.155 92.194 40.1 0.07833797 787.7948 308.0 4.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 299.9 12.6 (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 302.6 3.6 ■D (omit #3,4,5,11) CD Wtd mean = 302.0 1.5 (omit #3,4,5,11) C/) C/) isochron = 300.5 2.8 CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci' AS-1, C02 laser fusion single crystal sanidine, J = 0.00023788 ± 0.31% 4 amu discrimination = 1.01951 ± 0.13%, 40/39K = 0.0001 ± 100.0%, 36/37Ca : 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar739ArK Age (ka) Is.d. 1 1600 6 0.017 1.049 0.366 28.837 25.560 90.8 0.13447796 810.3118 318.0 4.0 3 3" 2 1600 6 0.035 1.319 0.357 27.041 28.860 72.5 0.18032372 785.3601 309.0 5.0 CD 3 1600 6 0.031 1.635 0.404 31.446 29.894 77.9 0.19221376 749.5380 296.0 4.0 ■DCD 4 1600 6 0.024 1.467 0.489 38.387 34.287 87.2 0.14127731 801.9807 315.0 3.0 O 5 1600 6 0.055 2.107 0.659 52.179 52.381 74.1 0.14927811 782.5954 308.0 3.0 Q. C 6 1600 6 0.040 0.898 0.321 24.720 28.329 66.2 0.13429306 766.0392 302.0 4.0 a O 7 1600 6 0.018 1.693 0.380 29.495 26.236 90.3 0.2121988 810.3118 318.0 4.0 3 8 1600 6 0.025 1.540 0.488 38.822 34.959 87.4 0.14682911 785.3601 309.0 4.0 "O O O 9 1600 6 0.031 0.831 0.316 24.254 26.552 72.0 0.12681913 793.6635 312.0 5.0 10 1600 6 0.036 1.866 0.398 30.532 31.562 72.4 0.22622186 763.2852 301.0 3.0 11 1600 6 0.065 1.184 0.421 32.601 43.044 59.3 0.13442769 818.6567 321.0 4.0 CD Q. 12 1600 6 0.081 1.084 0.437 33.308 46.475 51.9 0.12046119 755.0323 298.0 4.0 13 1600 6 0.042 0.613 0.228 17.294 28.534 61.8 0.13119963 1063.8745 407.0 7.0 14 1600 6 0.033 1.790 0.268 21.014 22.081 63.6 0.31530612 651.6748 260.0 5.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 312.4 30.0 ■D CD (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 313.8 4.6 (omit# 3,6,10,12,13,14) C/) Wtd mean = 313.6 1.6 C/) (omit# 3,6,10,12,13,14) Isochron = 316.0 2.2 CD ■ D O Q. C g Q.

■D CD

C/) C/)

8 ci' WP-1, C02 laser fusion single crystal sanidine, J = 0.00023782 ± 0,31% 4 amu discrimination = 1.01951 ± 0.13%, 40/39K = 0.0001 ± 100.0%, 36/37Ca = 0.000397 ± 17.6%, 39/37Ca = 0.000783 ± 8.75%

Crystal T(C) t (min.) 36Ar 37Ar 38Ar 39Ar 40Ar %40Ar* Ca/K 40Ar*/39ArK Age (ka) 1s.d. 1 1600 6 0.047 1.064 0.475 37.841 41.010 70.6 0.10104973 796.6353 313.0 4.0 3 2 1600 6 0.048 1.601 0.444 34.607 36.326 65.9 0.16626114 708.7029 281.0 4.0 3" CD 3 1600 6 0.044 0.939 0.337 26.170 32.311 69.8 0.12894992 891.7883 347.0 6.0 4 1600 6 0.032 1.760 0.340 26.746 27.982 78.6 0.23649666 838.3936 328.0 7.0 ■DCD O 5 1600 6 0.383 1.088 0.417 27.262 132.511 17.7 0.14342736 942.8942 365.0 10.0 Q. 6 1600 6 0.055 1.140 0.335 26.151 35.667 63.4 0.15666748 903.1012 351.0 10.0 C a 7 1600 6 0.139 1.085 0.246 18.441 53.129 28.2 0.21145291 860.8071 336.0 8.0 3O 8 1600 6 0.090 1.930 0.490 37.296 52.249 55.5 0.18597759 821.6486 322.0 7.0 "O 9 1600 6 0.042 1.759 0.307 23.645 29.006 68.7 0.2673629 863.6158 337.0 7.0 O 10 1600 6 0.044 0.617 0.366 28.603 35.206 72.3 0.07752238 928.6469 360.0 7.0 11 1600 6 0.066 1.679 0.479 37.557 46.419 65.3 0.16066547 849.5879 332.0 4.0 CD Q. 12 1600 6 0.053 0.531 0.371 29.690 33.276 55.7 0.06427415 633.0312 253.0 3.0 13 1600 6 0.061 2.557 0.404 31.256 41.755 60.2 0.29401868 843.9877 330.0 2.0 Note: isotope beams in mV risd = released, error in age includes J error, all errors 1 sigma Mean ± s.d. = 327.3 29.8 (36Ar through 40Ar are measured beam intensities, corrected for decay in age calculations) Mean ± s.d. = 330.8 5.1 ■D (omit# 1,2,3,5,6,10,12) CD Wtd mean = 330.4 1.8

C/) (omit# 1,2,3,5,6,10,12) C/) isochron = 325.8 2.2 APPENDIX B

MAJOR AND TRACE ELEMENT

GEOCHEMISTRY DATA

172

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 1. Major and trace element analyses of representative Obsidian Creek and Roaring Mountain member lavas. Sample GH-2 PH-5 LD-1 GLM-2 GLM-2a GRM-1 GRM-1 a AS-1

SiOj 76.00 76.45 76.28 65.19 61.54 74.28 56.80 74.70

AI2 O 3 13.14 12.38 12.72 13.83 17.60 12.94 15.12 13.01 TiOz 0.04 0.06 0.05 0.77 0.56 0.16 1.25 0.16

FeO 1.09 1 . 2 1 1 . 0 1 5.15 5.26 1.70 8 . 2 2 1.81

MnO 0 . 0 2 0 . 0 2 0 . 0 2 0.09 0 . 1 2 0.03 0.13 0.04 CaO 0.19 0.39 0.39 4.53 4.62 0.79 7.96 0.57

MgO 0.01 0.00 0.00 2.97 2.23 0.24 5.64 0.08 K2 O 4.66 4.64 4.79 3.05 2.13 5.51 1.54 4.96 N ^ O 4.26 4.06 4.16 3.32 3.96 3.38 3.03 3.92

P2 O 5 0 . 0 1 0 . 0 1 0 . 0 1 0 . 1 0 0.26 0 . 0 2 0.15 0 . 0 2 LOI 0 3 6 0.32 0.25 0.14 n.a. n.a. n.a. 0.17 Total 99.78 99.54 99.68 99.14 98.27 99.05 99.84 99.44

La 15.55 28.12 29.78 49.20 36.28 6 6 . 8 6 27.29 60.07 Ce 42.76 63.05 68.51 90.30 60.77 122.27 50.64 136.39

Pr 7 3 2 8 . 0 1 1 0 . 0 1 6.32 13.26 5.89 12.67 Nd 14.57 32.88 33.60 38.71 24.03 50.64 24.65 48.53

Sm 5.01 11.15 11.29 9.27 4.37 1 2 . 0 2 6 . 2 0 11.06

Eu 0.03 0 . 1 1 0.13 1.03 1.35 0.78 1.42 1.16 Gd 5.45 12.41 12.82 9.08 3.55 11.35 6.35 9.04 Tb 1.13 2.36 2.48 1.55 0.50 1.98 1.08 1.59

Dy 7.55 14.63 15.68 9.52 2.87 1 2 . 1 1 6.33 9.46 Ho 1.63 2.97 3.21 1.94 0.56 2.44 1.26 1.82 Er 4.56 7.81 8.60 4.97 1.48 6.63 3.29 4.80

Tm 0.71 1 . 1 2 1 36 0.72 0 . 2 2 0.93 0.46 0.72 Yb 4.56 6.79 7.48 4.35 1.41 5.66 2.71 4.57

Lu 0 . 6 6 0.97 1.06 0.63 0 . 2 2 0.82 0.40 0 . 6 6

Ba 4.66 6.42 9.55 188.76 1305.49 402.43 301.46 675.55 Th 44.32 29.94 25.90 13.92 4.37 20.42 5.67 21.28 Nb 83.50 57.20 54.21 27.81 5.72 36.75 16.73 38.56 Y 45.52 83.05 89.79 51.30 15.02 65.72 33.21 44.39

H f 9.42 8.64 7.51 6 . 0 0 3.94 7.89 4.01 8.67 Ta 6.94 4.45 4.15 2.05 0.35 2.80 1.17 2.77 U 8.51 6.15 6.39 3.08 0.91 4.49 1.23 3.81 Pb 45.19 33.87 35.16 20.95 15.91 30.74 9.37 24.87 Rb 382.02 256.78 257.85 115.68 36.08 185.00 46.84 166.31 Cs 5.18 4.44 4.31 2.55 0.90 4.12 0.87 2.67 Sr 1.56 1.25 1.44 153.62 691.55 29.52 314.36 32.24 Sc 0.69 0.60 0.55 13.90 9.23 2.60 21.98 2.74 Zr 140.27 161.51 137.58 170.54 145.37 211.45 128.18 262.55

Ni 0.00 0.00 2.00 33.00 10.00 11.00 86.00 6.00

Cr 0.00 0.00 0.00 40.00 17.00 12.00 88.00 0.00

V 0.00 0.00 0.00 90.00 74.00 7.00 170.00 1 1 . 0 0

Ga 31.00 25.00 27.00 2 1 . 0 0 2 0 . 0 0 2 0 . 0 0 2 2 . 0 0 2 2 . 0 0

Cu 7.00 19.00 1 2 . 0 0 2 0 . 0 0 17.00 7.00 42.00 4.00

Zn 60.00 6 8 . 0 0 69.00 82.00 75.00 77.00 89.00 76.00 Major elements are unnormalized in wt. %. Total = the sum of major element analyses including LOI values. Trace elements in ppm. Major elements and Ni, Cr, V, Ga, Cu, and Zn by XRF, all other trace elements by ICP-MS. Not analyzed = n.a.

173

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 1. Major and trace element analyses continued. Sam ple A S -la AS-2a WP-1 CS-1 OC-5 GR-2 CC-7 RF-2 SiOz 55.05 52.91 75.03 76.09 76.32 77.17 76.04 76.47

AI2 O 3 15.30 15.58 12.54 12.28 12.19 1 2 . 2 1 12.70 12.34 TiOz 1.49 1.59 0.13 0.09 0.07 0.14 O i l 0.09 FeO 9.20 929 1.69 1.27 1.14 1.34 1.13 1.38

MnO 0.16 0 .16 0.04 0.03 0 . 0 2 0 . 0 2 0 . 0 2 0 . 0 2 CaO 7.58 8.96 0.49 0.57 0.42 0.48 0.61 0.15

M gO 6 . 1 0 6.58 0 . 0 0 0 . 1 1 0 . 0 0 0.04 0.0 6 0 . 0 0

KzO 1.46 1 . 1 0 4.89 4.85 4 .90 5.11 5.20 4.88 NazO 3.05 2.95 3.87 3.71 339 3.39 3.26 3.93

PzO; 0.19 0 . 2 0 0 . 0 2 0 . 0 1 0 . 0 1 0.01 0.01 0.02 LOI n.a. n.a. 0.13 0.50 0.56 0.15 0.42 0.24 Total 99 j? 99.81 98.82 99.52 99.33 100.05 99.55 99.52

La 76.87 42.88 63.75 61.17 59.57 70.03 6 9 .3 4 70.78 Ce 143.64 67.91 116.67 114.86 112.70 116.29 122.81 110.69

Pr 2 1 . 2 0 10.24 12.32 12.56 1 2 . 2 1 1 2 . 6 8 12.45 14.98 N d 84.40 43.98 46.20 47.96 46 .9 6 45.83 43 .6 7 59.05 Sm 20.49 11.47 10.19 12.34 11.96 9.70 9.47 16.15 Eu 1.73 1.77 1.03 0.19 0.16 0.49 0.4 0 0.54 Gd 18.06 12.71 9.00 12.43 12.15 8.54 8.81 16.93

Tb 3.12 2 . 1 0 1.54 2.28 2 . 2 1 1.42 1.57 3.16 D y 18.19 12.73 9.22 14.21 13.70 8 j# 9.63 19.92 Ho 3.32 2.46 1.85 238 2.79 1.65 1.98 4 .06 Er 839 6.16 4.80 7.65 7.57 4.3 9 5.47 10.97

Tm 122 0.84 0.70 1 . 1 2 1.08 0.62 0.8 2 1.58 Yb 7.45 4.7 9 438 636 6.64 3.78 5.07 9.31 Lu 1.07 0.67 0.65 0.99 0.95 0.56 0.75 1.32

Ba 41 6 .6 6 36 7 .0 0 583.27 38.79 34.43 287.49 228.89 75.36 Th 6.75 3.62 20.15 29.94 29.32 20 .9 0 3 3 .4 4 24.11 N b 18.94 15.46 39.06 44.17 43.15 24.93 2 1 .7 0 66.94 Y 73.17 61.57 46.92 78.76 76.99 43.21 56.25 100.99 H f 4.74 4.2 0 8.65 7.52 7.28 6.54 5.62 10.91 Ta 1.23 1.38 236 3.47 3.40 1.90 2.1 4 4.71 U 2.19 1.25 3.78 6.60 6.44 3.79 5.30 4.67 Pb 25.09 7.28 28.91 35.54 35 .3 0 24 .4 4 30.08 47.11

Rb 43.41 2 5 .3 0 170.08 226.58 223.92 147.47 2 0 2 . 0 0 208.54

Cs 0.96 0.49 3.46 4.92 4.82 1.81 2 .99 2 . 0 2 Sr 273.67 348.00 24.86 7.73 2.35 15.18 15.59 2.47

Sc 25.10 2 8 .2 0 2 . 1 0 1.81 1.29 1.71 2 . 1 0 0.49 Zr 159.72 148.00 251.93 157.31 153.75 176.30 130.80 251.25

N i 1 0 1 . 0 0 109.00 1 . 0 0 1 . 0 0 0 . 0 0 3.00 1 . 0 0 1 . 0 0

Cr 135.00 140.00 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0

V 187.00 2 1 7 .0 0 4.00 8 . 0 0 1 . 0 0 4.00 0 . 0 0 0 . 0 0

Ga 2 0 . 0 0 2 0 . 0 0 24.00 24.00 24.00 2 0 . 0 0 19.00 27.00

Cu 33.00 36.00 1.00 0.00 10.00 2 . 0 0 3 .00 1 2 . 0 0

Zn 2 50.00 1 2 0 . 0 0 77.00 6 8 . 0 0 73 .0 0 4 7 .0 0 33 .0 0 163.00

174

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CD ■ D O Q. C g Q.

■D CD

C/) 3o" O Table 2. Major and trace element XRF instrumental precision.

Unnormailized Major SiOz AI2 O3 TiOz FeO MnO CaO MgO KzO NazO P2 O5 Total 8 Standard Deviation (Wt. %) 0.18 0 .1 1 0.004 0 .0 1 0 .0 0 1 0 .0 1 0.1 0.09 0.05 0.003 0.36 ci'

Normalized Major SiOz AI2O3 TiOz FeO MnO CaO MgO KzO NazO P2O5

Standard Deviation (Wt. %) 0.09 0.07 0.004 0 .0 1 0 .0 0 1 0 .0 1 0.1 0.07 0.05 0 . 0 0 2

3 Trace Elements Ni Cr Sc V Ba Rb Sr Zr Y Nb Ga 3" CD Standard Deviation ( ppm) 1 2 2 5 9 1 1 1 1 0.5 1 ■DCD O Q. Cu Zn Pb La Ce Th C a O 2 2 2 10 10 2 ^ Precision is based on Washington State University in-house standard (GSP-1) measured over an 8 month period.

CD Q.

Table 3. Trace element ICP-MS instrumental precision ■D CD Element La Ce Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu ± 2 a &98 1.24 0.96 0.3 0.08 0.16 0 .0 2 0 .2 0.04 0 .1 2 0 . 0 2 0.06 0 .0 2 C/) C/) Element Ba Th Y Hf Ta U Pb Rb Cs Sr Sc Zr ± 2 a 25 0.98 0.58 0.14 0.04 0 .2 2 &58 1.34 0.06 n.d. n.d. n.d. Precision is based on Washington State University in-house standard (BCR-P) measured over a 4 month period (September - December, 1994). No data (n.d.) were reported on Sr, Sc, or Zr. CD ■ D O Q. C 8 Q.

■D CD

C/) C/)

Table 4. Loss on ignition calculations for the extracaldera rhyolites CD Crucible + Sample Sample wt. Wt. After Wt. Lost after LOI for Wt after Wt. Lost LOI for Total wt. % Total Sample wt. (g) wt. (g) (g) 110° C 110° C 110° C 950° C after 950° C 950° C Loss (g) LOI 3. GH-2 46.8645 49.3472 2.4827 49.3459 0.0013 0.052 49.3382 0.0077 0.310 0.0090 0.363 3" CD PH-5 44.9424 47.2627 2.3203 47.2613 0.0014 0.060 47.2552 0.0061 0.263 0.0075 0.323 ■DCD LD-1 44.0924 46.7886 2.6962 46.7875 0.0011 0.041 46.7819 0.0056 0.208 0.0067 0.248 O GLM-2 44.8526 47.2747 2.4221 47.2727 0.0020 0.083 47.2732 0.0015 0.062 0.0035 0.145 CQ. a GR-2 44.5411 47.6108 3.0697 47.6094 0.0014 0.046 47.6062 0.0032 0.104 0.0046 0.150 o 3 -J OC-5 48.1829 50.6133 2.4304 50.6094 0.0039 0.160 50.5996 0.0098 0.403 0.0137 0.564 ■D G h O CS-1 44.4864 46.7096 2.2232 46.7069 0.0027 0.121 46.6984 0.0085 0.382 0.0112 0.504 W P-1 45.0690 47.6320 2.5630 47.6306 0.0014 0.055 47.6287 0.0019 0.074 0.0033 0.129 CD 47.3584 49.7137 0.0024 0.102 49.7037 0.0076 0.323 0.0100 0.425 Q. CC-7 2.3553 49.7113 RF-2 44.1697 46.7570 2.5873 46.7564 0.0006 0.023 46.7509 0.0055 0.213 0.0061 0.236 AS-1 45.5692 48.0721 2.5029 48.0718 0.0003 0.012 48.0681 0.0037 0.148 0.0040 0.160 ■D AS-1* 44.4444 46.7587 2.3143 46.7573 0.0014 0.060 46.7547 0.0026 0.112 0.0040 0.173 CD Note: With the exception of OC-5, CS-1, and CC-7 glasses, all other samples C/) C/) are porphyritic rhyolite. AS-1* represents a duplicate run for precision. APPENDIX C

SAMPLE OUTLIERS: THE COUGAR CREEK AND RIVERSIDE

RHYOLITES AND THE GRIZZLY LAKE ANDESITE

177

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Cougar Creek and Riverside Rhyolites

The originally defined Roaring Mountain member rhyolites include the Cougar Creek

dome and Riverside flow which erupted near the Madison River -20 - 25 km west of the

Norris-Mammoth corridor. New ""^Ar/^^Ar ages (Chapter 6 ) have determined that the

Cougar Creek (CC-7, 358 ka) and Riverside (RF-2, 526 ka) rhyolites are considerably

older than the other Roaring Mountain rhyolites (118 - 80 ka). The sparsely porphyritic

nature of the Cougar Creek and Riverside rhyolites is also not typical of the aphyric

Roaring Mountain rhyolites with the exception of the Gibbon River flow, which is

aphyric to porphyritic (Chapter 5). The ages and petrography of the Cougar Creek and

Riverside rhyolites suggest that they are more likely related to the porphyritic Obsidian

Creek member rhyolites (326 - 134 ka), which closely post-date their eruption.

Trace element and isotopic geochemistry, however, shows that the Cougar Creek and

Riverside rhyolites are unrelated to either member (Chapter 7). Samples RF-2 and CC-7

both overlap the Roaring Mountain member rhyolites in LREE’s however RF-2 is

enriched in HREE’s whereas CC-7 plots between the Gibbon River (GR-2, 118 ka) and

the indistinguishable Obsidian Cliff (OC-5, 106 ka) and Crystal Spring (80, ka) rhyolites

(Figure 7.07). CC-7 and RF-2 also plot off all age versus trace element trends (Figures

7.04 and 7.05) suggesting that they are unrelated to either the Roaring Mountain or the

Obsidian Creek member.

Finally, Nd and Pb isotopic ratios provide further evidence that the Cougar Creek and

Riverside rhyolites are distinct. Both overlap the other rhyolites in ^^Sr/^Sn, and RF-2

overlaps in SNd at -13.1, whereas CC-7 is distinctly lower at -15.8 (Table 7.01 and Figure

7.09). CC-7 and RF-2 are less radiogenic in their Pb isotopic compositions and therefore

178

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. cluster well outside the main group of extracaldera rhyolites in and

208pb/204pb yersus plots (Table 7.01 and Figure 7.11) indicating they

originated from sources distinct from the Roaring Mountain and Obsidian Creek member

rhyolites.

The above evidence indicates that the Cougar Creek and Riverside rhyolites are

unrelated to either the Roaring Mountain member as originally defined, or to the

Obsidian Creek member. These units are therefore not discussed further in the evolution

of the extracaldera rhyolites.

The Grizzly Lake Andésite

Sample GLM-2a is an andésite enclave (Figure 7.01) that was collected from the

Grizzly Lake mingled lava and selected as a sample thought to represent the abundant

mafic clots/enclaves within the dacite host flow. However, pétrographie analysis shows

that the mineralogy of sample GLM-2a is distinct from mafic clots exposed in a thin

section of the dacitic host (sample GLM-2). Mafic clots (Figure 5.05) are mostly

composed of glomerocrysts of plagioclase and clinopyroxene set in a holocrystalline

groundmass of the same minerals, and are similar to the mineralogy of basaltic-andesite

enclaves from the Gardner River and Appolinaris Spring mingled lavas (Table 5.01).

However, GLM-2a is a complex assemblage of plagioclase, quartz, biotite, sanidine,

clinopyroxene and hornblende phenocrysts (Table 5.01) set in a holocrystalline

plagioclase groundmass. The presence of zoned and embayed quartz xenocrysts, and

zoned plagioclase and sanidine phenocrysts indicate that this andésite enclave was in

disequilibrium with the host magma.

179

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A recent study on the petrogenesis of andésite and dacite lavas that erupted between

55-52 Ma from Washburn Volcano, a major Eocene calc-alkaline volcanic center

exposed in the northeastern region of Yellowstone as part of the Absorka Volcanic

Province (Feeley, et al., 2002) suggests an origin for GLM-2a. Nd and Sr isotopic

compositions and trace element data indicate that Washburn magmas were derived from

deep crustal rocks similar in composition to granulite xenoliths of the Wyoming

Province. GLM-2a is similar to Washburn andésites in its major element, trace element,

and isotopic geochemistry.

Major element compositions of GLM-2a are within 1.0 wt. % of those reported for

the Washburn andésites. Chondrite-normalized REE and MORB-normalized trace

element plots (Figures 1 and 2) show the overlap of GLM-2a with the Washburn

andésites. The MORB-normalized trace element plot shows that GLM-2a and Washburn

andésites are characterized by strong enrichments in LIL’s (Sr, K, Rb, Ba, and Th) and

depletions in HFSE’s (Ta, Nb and Ti), which are signatures diagnostic of subduction-

derived magmas. The Sr and Nd isotopic composition of GLM-2a is also

indistinguishable from Washburn andésites, which have ^’Sr/^^Sq values that range

0.7058-0.7071 and '"'^Nd/’'^"'Ndi that cluster at 0.5114 (Feeley et al. 2002). GLM-2a is

therefore interpreted to represent, not a crustal xenolith, but a Washburn andésite xenolith

brought to the surface by the hybrid Grizzly Lake dacite.

The volcanic texture of the GLM-2a enclave is consistent with its derivation as a

xenolith of the Washburn volcanics. The disequilibrium mineral textures observed in

sample GLM-2a may have been produced by reheating during transport in Swan Lake

Flat magma. The possibility that Swan Lake Flat basalt ascended through Washburn

180

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. andésite flows is very likely considering the close proximal location of Washburn

Volcano (~ 20 km east of Norris-Mammoth corridor) and the fact that these volcanic

rocks are exposed ~ 6 km southeast of the Grizzly Lake flow. An interpretive

reconstruction of the Yellowstone Plateau before the onset of plateau volcanism around 2

Ma suggests that lavas of the Absorka Volcanic Province extended as far west as the

Norris-Mammoth corridor (Christiansen, 2001).

181

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. / s . GLM-2a andésite ^ Mt. Washburn andésites 100 w

o o 0:

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 1. Chondrite-normalized REE plot (Sun and McDonough, 1989) showing that andésite enclave GLM-2a (from the Grizzly Lake mingled lava) is indistinguishable from the Mt. Washburn andésites (Feeley et ah, 2002) in rare earth element abundances.

100

GLM-2a andésite ^ Mt. Washburn andésites

10 DU O

Ü o DC 1

1 Sr K Rb Ba Th Ta Nb Ce P Zr Hf Sm Ti Y Yb

Figure 2. Morb-normalized (Pearce, 1983) trace element plot showing that trace element abundances from an andésite enclave GLM-2a (from the Grizzly Lake mingled lava) are indistinguishable from Mt. Washburn andésites (Feeley et al., 2002).

182

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES

Anders, M.H., and Sleep, N.H., 1992, Magmatism and extension; The thermal and mechanical effects of the Yellowstone hotspot: Journal of Geophysical Research, v. 97,p . 15379-15393.

Anders, M.H., Geissman, J.W., Piety, L.A., and Sullivan, J.T., 1989, Parabolic distribution of circumeastem Snake river Plain seismieity and latest Quaternary faulting: Migratory pattern and association with the Yellowstone hotspot: Journal of Geophysical Research, v. 94, no. B2, p. 1589-1621.

Annen, C., and Sparks, R.S.J., 2002, Effects of repetitive emplacement of basaltie intrusions on thermal evolution and melt generation in the crust: Earth and Planetary Science Letters, v. 203, p. 937-955.

Bachmann, O., and Bergantz, G.W., 2004, On the origin of crystal-poor rhyolites: Extraeted from batholithic crystal mushes: Journal of Petrology, v. 45, no. 8 , p. 1565- 1582.

Bachmann, O., Dungan, M.A., and Lipman, P.W., 2002, The Fish Canyon magma body, San Juan volcanic field, Colorado: Rejuvenation and eruption of an upper-crystal batholith. Journal of Petrology, v. 43, p. 1469-1503.

Bennett, K., in progress [M.S. thesis]: University of Nevada, Las Vegas.

Bindeman, I.N., and Valley, J.W., 2000, The formation of low-0rhyolites after caldera collapse at Yellowstone, Wyoming, USA: Geology, v. 28, p. 719-722.

Bindeman, I.N., and Valley, J.W., 2001, Low- 6 '^0 rhyolites from Yellowstone: Magmatic evolution based on analyses of zireons and individual phenocrysts. Journal of Petrology, v. 42, p.1491-1517.

Bindeman, I.N., Valley, J.W., Wooden, J.L., Persing, H.M., 2001, Post-caldera voleanism: In situ measurement of U-Pb age and oxygen isotope ratio in Pleistocene zircons from Yellowstone caldera: Earth and Planetary Science Letters, v. 189, p. 197 206.

Camp, V.E., 1995, Mid-Miocene propagation of the Yellowstone mantle plume head beneath the Columbia River Basalt souree region: Geology, v. 23, p. 435-438.

183

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Cebula G.T., Kunk, M.J., Mehnert, H.H., Naeser, C.W., Obradovich, J.D., and Sutter, J.F., 1967, The Fish Canyon Tuff, a potential standard for the "**^Ar/^^Ar and fission-

track dating methods: Terra Cognita, v. 6 , no. 2, p. 139-140.

Christensen, J.N., Halliday, A.N., 1996, Rb-Sr ages and Nd isotopic compositions of melt inclusions from the Bishop Tuff and the generation of silicic magma: Earth and Planetary Science Letters, v. 144, p. 547-561.

Christiansen, R.L., 2001, The Quaternary and Pliocene Yellowstone Plateau volcanic field of Wyoming, Idaho and Montana: U.S. Geological Survey Professional Paper, 729-B, 145p.

Christiansen R.L. and Blank, H R., Jr., 1972, Volcanic stratigraphy of the Quaternary rhyolite plateau in Yellowstone National Park: U.S. Geological Survey Professional Paper 729-B, 18p.

Christansen, R.L. and Yeats, R.S., 1992, Post-Laramide geology of the U.S. Cordilleran region, in Burchfiel, B.C., Lipman, P.W., and Zoback, M.L., eds.. The Cordilleran orogen: Conterminous U.S.: Boulder, Co., Geological Society of America, The Geology of North America, v. G-3, p. 261-406.

Christiansen, R.L., Foulger, G.R., Evans, J.R., 2002, Upper-mantle origin of the Yellowstone hotspot: Geological Socitey of America Bulletin, v. 114, p. 1245-1256.

Davies, G.R., and Halliday, A.N., 1998, Development of the Long Valley rhyolitic magma system: Strontium and neodymium isotope evidence from glasses and individual phenocrysts: Geochim. Cosmochim. Acta, v. 62, p. 3561-3574.

Davies, G.R., Halliday, A.N., Mahood, G.A., Hall, C.M., 1994, Isotopic constraints on the production rates, crystallization histories and residence times of pre-caldera silicic magmas. Long Valley, California: Earth and Planetary Science Letters, v. 125, p. 17- 37.

Depaolo, D.J., 1981, Trace element and isotopic effects of combined wall-rock assimilation and fractional crystallization: Earth and Planetary Science Letters, v. 53, p. 189-202.

Depaolo, D.J. and Manga, M., 2003, Deep Origin of Hotspots: The mantle plume model: Science, v. 300, p. 920-921.

184

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Doe, B.R., Leeman, W., Christiansen, R.L., Hedge, C.E., 1982, Lead and strontium isotopes and related trace elements as genetic tracers in the upper Cenozoic rhyolite basalt association of the Yellowstone Plateau volcanic field: Journal of Geophysical Research, v. 87, p. 4785-4806.

Dudas, P.O., Carlson, R.W., Eggler, D.H., 1987, Regional Middle Proterozoic enrichment of the subcontinental mantle source of igneous rocks from central Montana: Geology, v. 15, p. 22-25.

Faure, G., 1986, Principles of isotope geology: New York, J. Wiley and Sons, Inc., 589 p.

Feeley, T.C., Cosca, M.A., and Lindsay, C.R., 2002, Petrogenesis and implications of calc-alkaline cryptic hybrid magmas from Washburn Volcano, Absaroka Volcanic Province, USA: Journal of Petrology, v. 43, no. 4, p. 663-703.

Fenner, C.N., 1938, Contact relations between rhyolite and basalt on the Gardiner River, Yellowstone Park: Geological Socitey of America Bulletin, v. 49, p. 1441-1484.

Fenner, C.N., 1944, Rhyolite-basalt complex on Gardiner River, Yellowstone Park, Wyoming: A discussion: Geological Socitey of America Bulletin, v. 55, p. 1081- 1096.

Fournier, R.O., 1989, Geoehemistry and dynamics of the Yellowstone National Park hydrothermal system: Annual Reviews of Earth and Planetary Science, v. 17, p. 13- 53.

Fournier, R.O., and Pitt, A.M., 1985, The Yellowstone magmatic-hydrothermal system, U.S.A.: in Stone, C., ed.. Geothermal Resources Council International Symposium on Geothermal Energy, International Volume, p. 319-327.

Fraser, G.D., Witkind, I.J., and Nelson, W.H., 1964, A geological interpretation of the epicentral area - the dual basin concept, in The Hebgen Lake, Monatna, earthquake of August 17, 1959: U.S. Geological Survey Professional Paper 435-F, p. 31-35.

Friedman, I., Lipman, P.W., Obradovich, J.D., Gleason, J.D., Christiansen, R.L., 1974, Meteoric water in magmas: Science, v. 184, p. 1069-1072.

Geist, D., and Richards, M., 1993, Origin of the Columbia Plateau and Snake River Plain: Deflection of the Yellowstone plume: Geology, v. 21, p. 789-792.

185

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Halliday, A.N., 1990, Reply to comment of R.S.J. Sparks, H.E. Huppert, and CJ.N. Wilson on “Evidence for long residence times of rhyolitic magma in the Long Valley magmatic system: the isotopic record in precaldera lavas of Glass Mountain”: Earth and Planetary Science Letters, v. 99, p. 390-394.

Halliday, A.N., Mahood, G.A., Holden, P., Metz, J.M., Dempster, T.J., and Davidson, J.P., 1989, Evidence for long residence times of rhyolitic magma in the Long Valley magmatic system: the isotopic record in precaldera lavas of Glass Mountain: Earth and Planetary Science Letters, v. 94, p. 274-290.

Hawkes, L., 1945, The Gardiner River rhyolite-basalt complex: Geological Magazine, V. 82, p. 182-184.

Hawkesworth, C.J., Blake, S., Evans, P., Hughes, R., Macdonald, R., Thomas, L.E., Turner, S.P., Zellmer, G., 2000, Time scales of crystal fractionation in magma chambers - Integrating physical, isotopic and geochemical perspectives: Journal of Petrology, v. 41, no.7, p. 991-1006.

Henry, P.A., Mueller, J.L., Wooden, J.L., Warner, and R., Le Berman, 1982, Granulite grade supercrustal assemblages of the Quad Creek area, eastern Beartooth Mountains, Montana, in Precambrian Geology of the Beartooth Mountains, Mueller, P., and Wooden J., eds., Montana Bureau of Mines Geology Special Publication, v. 84, p. 147-156.

Heumann, A, Davies, G.R., Elliott, T., 2002, Crystallization history of rhyolites at Long Valley, California, inferred from combined U-series and Rb-Sr isotope systematics: Geochem.Cosmochem. Acta, v. 6 6 , p. 1821-1837.

Hildreth, W., 1981, Gradients in silicic magma systems: Implications for lithospheric

magmatism: Journal of Geophysical Research, v. 8 6 , p. 10153-10192.

Hildreth, W., Christiansen, R.L., O'Neil, J.R., 1984, Catastrophic isotopic modification of rhyolitic magma at times of caldera subsidence, Yellowstone Plateau volcanic field: Journal of Geophysical Research, v. 89, p. 8339-8369.

Hildreth, W., Halliday, A.N., Christiansen, R.L., 1991, Isotopic and chemical evidence concerning the genesis and contamination of basaltic and rhyolitic magma beneath Yellowstone Plateau volcanic field: Journal of Petrology, v. 32, p. 63-138.

Humphreys, E D., and Dueker, K.G., 1994, Western U.S. upper mantle strueture: Journal of Geophysical Research, v. 99, no. B5, p. 9615-9634.

186

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Humphreys, E.D., and Dueker, K.G., 1994, Physical state of the western U.S. upper mantle: Journal of Geophysical Research, v. 99, no. B5, p. 9635-9650.

Humphreys, E.D., Dueker, K.G., Schutt, D.L., and Smith, R.B., 2000, Beneath Yellowstone: Evaluating plume and nonplume models using teleseismic images of the upper mantle: GSA Today, v. 10, p. 1-7.

Huppert, H.E., Sparks, R.S.J., 1988, The generation of granitic magmas by intrusion of basalt into continental crust: Journal of Petrology, v. 29, p. 599-624.

Husen, S., Smith, R.B., Waite, G.P., 2003, Evidence for gas and magmatic sources beneath the Yellowstone volcanic field from seismic tomographic imaging: Journal of Volcanology and Geothermal Research, v. 131, p. 397-410.

Iddings, J.P., 1899, in Hague, A., and others. The rhyolites. Geology of the Yellowstone National Park: U.S. Geological Survey Monograph 32, Part 2, 893 p.

Johnson, D.M., Hooper, P R., and Conrey, R.M., 1999, XRF analysis of rocks and minerals for major and trace elements on a single low dilution Li-tetraborate fused bead: JCPDS International Centre for Diffraction Data, v. 41, p. 843-867.

Justet, L., Spell, T.L., 2001, Effusive eruptions from a large silicic magma chamber: the Bearhead Rhyolite, Jemez volcanic field, NM: Journal of Volcanology and Geothermal Research, v. 107, p. 241-264.

Kharaka, Y.K., Sorey, M.L., Thordsen, J.J., 2000, Large-scale hydrothermal fluid discharges in the Norris-Mammoth corridor, Yellowstone National Park, USA: Journal of Geochemical Exploration, v. 69-70, p. 201-205.

Knaack, C., Cornelius, S., and Hooper, P., 1994, Trace element analyses of rocks and minerals by ICP-MS: Washington State University, Geo Analytical Lab webpage, http://www.wsu.edu:8080/~geology/Pages/Services/ICP.html.

Krogh, T. E., 1973, A low contamination method for hydrothermal decomposition of zircon and extraction of U and Pb for isotopic age determinations: Geochimica et Cosmochimica Acta, v. 37, p. 485-494.

Krogh, T.E., 1982, Improved accuracy of U-Pb ages by the creation of more concordant systems using an air abrasion technique: Geochimica et Cosmochimica Acta, v. 46, p. 637-649.

Langmuir, C H., Vocke, R.D., Hanson, G.N., and Hart, S.R., 1978, A general mixing equation with applications to Icelandic basalts: Earth and Planetary Science Letters, v. 37, p. 380-392.

187

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Leeman, W.P., Menzies, M.A., Matty, D.J., and Embree, G.F., 1985, Strontium, neodymium and lead isotopic compositions of deep crustal xenoliths from the Snake River Plain: evidence for Arhean basement: Earth and Planetary Science Letters, v. 75, p. 354-368.

Lanphere, M.A., Champion, D.E., Christiansen, R.L., Izett, G.A., Obradovich, J.D., 2002, Revised ages for tuffs of the Yellowstone Plateau volcanic field: Assignment of the Hucklebery Ridge Tuff to a new geomagnetic polarity event: Geological Society of America Bulletin, v. 114, p. 559-568.

Love, J.D., and Keefer, W.R., 1975, Geology of sedimentary rocks in southern Yellowstone National Park, Wyoming, in The Geology of Yellowstone National Park: U.S.: Geological Survey Professional Paper 729-D, 60 p.

Love, J.D., Reed, J.C., Jr., Christiansen, R.L., and Stacy, J R., 1973, Geologic block diagram and tectonic history of the Teton region: U.S. Geological Survey Miscellaneous Geologic Investigations Map 1-730.

Ludwig, K.R., 1998, Users Manual for Isoplot/Ex rev. 2.49: Berkely Geochronology Center, Special Publication No. 1, 56 p.

Mahood, G. A., 1990, Second reply to comment of R.S.J. Sparks, H.E. Huppert and C.J.N. Wilson on “Evidence for long residence times of rhyolitic magma in the Long Valley magmatie system: the isotopie record in the precaldera lavas of Glass Mountain”: Earth and Planetary Science Letters, v. 99, p. 395-399.

Mahood, G. A., and Halliday, A. N., 1988, Generation of high-siliea rhyolite: A Nd, Sr, and O isotopic study of Sierra La Primavera, Mexican Neovoleanie Belt: Contributions to Mineralogy and Petrology, v. 100, p. 183-191.

Mankinen, E.A., Gromme, C.S., Dalrymple, G.B., Lanphere, M.A., and Bailey, R.A., 1986, Paleomagnetism and K/Ar ages of volcanic rocks from Long Valley Caldera, California: Journal of Geophysical Research, v. 91, p. 633-652.

Manley, C.R., Bacon, C.R., 2000, Rhyolite thermobarometry and the shallowing of the magma reservoir, Coso volcanic field, California: Journal of Petrology, v. 41, p. 149-174.

Miller, D.S., Smith, R.B., 1999, P and S velocity structure of the Yellowstone volcanic field from local earthquake and eontrolled-source tomography: Journal of Geophysical Research, v. 104, p. 15105-15121.

McDougall, I., and Harrison, T.M., 1999, Geochronology and Thermochronology by the ""^Ar/^^Ar Method, 2"^ edition, Oxford University Press, New York, 269 p.

188

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Murphy, M.D., Sparks, R.S.J., Barclay, J., Carroll, M.R., and Brewer, T.S., 2000, Remobilization of andesitic magma by intrusion of mafic magma at the Soufiriere Hills volcano, Montserrat, West Indies: Journal of Petrology, v. 41, p. 21-42.

Myers, W.B., and Hamilton, W., 1964, Deformation accompanying the Hebgen Lake earthquake of August 17, 1959,in The Hebgen Lake, Montana, earthquake of August 17, 1959: U.S.: Geological Survey Professional Paper 435-1, p. 55-98.

Obradovich, J.D., 1992, Geochronology of the late Cenozoie volcanism of Yellowstone National Park and adjoining areas, Wyoming and Idaho: U.S.: Geological Survey Open-file Report 92-408, 45 p.

Pardee, J.T., 1950, Late Cenozoie block faulting in western Montana: Geological Society of America Bulletin, v. 61,p. 359-406.

Parsons, T., Thompson, G.A., and Sleep, N.H., 1994, Mantle plume influence on the Neogene uplift and extension of the U.S. western Cordillera?: Geology, v. 22, p. 83- 86 .

Patchett, P.J., and Ruiz, J., 1987, Nd isotopie ages of crust formation and metamorphism in the Precambrian of eastern and southern Mexico: Contributions to Mineralogy and Petrology, v. 96, p. 523-528.

Perkins, M.E. and Nash, B.P., 2002, Explosive silicic volcanism of the Yellowstone hotspot: The ash fall tuff record: Geological Society of America Bulletin, v. 114, p. 367-381.

Pierce, K.L. and Morgan, L.A., 1992, The track of the Yellowstone hot spot: Volcanism, faulting, and uplift, in Link, P.K., Kuntz, M.A., and Platt, L.B., eds.. Regional Geology of eastern Idaho and western Wyoming: Geological Society of America Memoir 179, p. 1-53.

Reid, M R., and Coath, C D., 2000, In situ U-Pb ages of zircons from the Bishop Tuff: No evidence for long crystal residence times: Geology, v. 28, p. 443-446.

Reid, M R., Coath, C D., Harrison, T.M., McKeegan, K.D., 1997, Prolonged residence times for the youngest rhyolites associated with Long Valley Caldera: ^^^Th-^^^U ion microprobe dating of young zircons: Earth and Planetary Science Letters, v. 150, p. 27-39.

Ruppel, E.T., 1972, Geology of pre-Tertiary rocks in the northern part of Yellowstone National Park, Wyoming, in Geology of Yellowstone National Park: U.S. Geological Survey Professional Paper 729-A, 6 6 p.

Smith R.B., and Braile, L.W., 1994, The Yellowstone hotspot: Journal of Volcanology and Geothermal Research, v. 61, p. 121-187.

189

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Smith R.B., Shuey, R.T., Pelton, J.R., and Bailey, J.P., 1977, Yellowstone hot spot: Contemporary tectonics and crustal properties from earthquake and aeromagnetie data: Journal of Geophysial Research, v. 82, p. 3665-3676.

Smith, R.L., 1979, in Chapin, C.E., Elston, W.E., eds.. Ash-flow magmatism: GSA Spee. Paper v. 180, p. 5-27.

Smith, R.L., and Bailey, R.A., 1968, Resurgent cauldrons, in Coats, R.R., Hay, R.L., and Anderson, C.A., eds.. Studies in volcanology - A memoir in honor of Howell Williams: GSA Memoir 116, p. 613-662.

Sparks, R.S.J., Huppert, H E., Wilson, C.J.N., 1990, Comment on “Evidence for long residence times of rhyolitic magma in the Long Valley magmatie system: the isotopic record in precaldera lavas of Glass Mountain”: Earth and Planetary Science Letters, v. 99, p. 387-389.

Spell, T. L., and Nastanski, N. M., 2004, Ion Microprobe ^"^Pb/^^^U and Zircon Ages From Extracaldera Rhyolites at Yellowstone: Constraints on Magma Residence Times and Evolution: Geological Society of America Abstracts with programs, v. 36, no. 5, p. 431-432.

Spell, T.L., Kyle, P R., Thirlwall, M.F., Campbell, A.R., Isotopie and geochemical constraints on the origin and evolution of postcollapse rhyolites in the Valles Caldera, New Mexico: Journal of Geophysical Research, v. 98, p. 19,723-19,739.

Steiger, R.H., and Jager, E., 1977, Subeommission on geochronology: Convention on the use of decay constants in geo- and cosmo-chronology: Earth and Planetary Science Letters, v. 36, p. 359-362.

Steven, T.A., Mehnert, H.H., and Obradovich, J.D., 1967, Age of volcanic activity in the San Juan Mountains, Colorado: U.S. Geological Survey Professional Paper, p. D47- D55.

Struhsacker, D.W., 1978, Mixed basalt-rhyolite assemblages in Yellowstone National Park: The petrogenetic significance of magma mixing [M.S. thesis]: Missoula, University of Montana, 112 p.

Vasquez, George A., and Reid, Mary R., 2002, Time scales of magma storage and differentiation of voluminous high-silica rhyolites at Yellowstone caldera, Wyoming: Contributions to Mineralogy and Petrology, v. 144, p. 274-285.

Watson, E.B., and Harrison, T.M., 1983, Zircon saturation revisited: Temperature and composition effects in a variety of crustal magma types: Earth and Planetary Science Letters, v. 64.

Wendt, I., and Carl, C., 1991, The statistical distribution of the mean squared weighted

190

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. deviation: Chemical Geology, v. 86, p. 275-285.

Wiebe, R.A., Manon, M.R., Hawkins, D.P., McDonough, W.F., 2004, Late-stage mafic injection and thermal rejuvenation of the Vinalhaven Granite, Coastal Maine: Journal of Petrology, v. 45, p. 2133-2153.

Wilcox, R.E., 1944 Rhyolite-basalt complex on Gardiner River, Yellowstone Park, Wyoming: Geological Socitey of America Bulletin, v.55, p. 1047-1079.

Wooden, J.L., and Mueller, P. A., 1988, Pb, Sr, and Nd isotopie compositions of a suite of Late Archean, igneous rocks, eastern Beartooth Mountains: implications for crust- mantle evolution: Earth and Planetary Science Letters, v. 87, p. 59-72.

Wooden, J.L., Hunt, D.K, Bowes, D R., 1982, Geochemistry and Rb-Sr geochronology of Archean rocks from the interior of the southeastern Beartooth Mountains, Montana and Wyoming: Montana Bureau of Mines Geol. Spee. Publ., v. 84, p. 45-56.

Wolff, J.A., Balsley, S.D., Gregory, R.T., 2001, Oxygen isotope disequilibrium between quartz and sanidine from the Bandelier Tuff, New Mexico, consistent with a short residence time of phenocrysts in rhyolitic magma: Journal of volcanology and Geothermal Research, v. 116, p. 119-135.

Wolff, J.A., and Gardner, J.N., 1995, Is the Valles ealdera entering a new cycle of activity?: Geology, v. 23, p. 411-414.

Woods, A.W., and Huppert, H E., 2003, On magma chamber evolution during slow effusive eruptions: Journal of Geophysical Research, v. 108, no. B8.

Zellmer, G.F., Sparks, R.S.J., Hawkesworth, C.J., Wiedenbeck, M., 2003, Magma emplacement beneath Montserrat: Insights from Sr and Ba zonation in plagioclase phenocrysts: Journal of Petrology, v. 44, no. 8, p. 1413-1431.

191

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. VITA

Graduate College University of Nevada, Las Vegas

Nicole Marie Nastanski

Home Address: 1453 Lorilyn Ave., #4 Las Vegas, NV 89119

Degrees: Bachelor of Science, Geology, 2002 University of Pittsburgh

Publications:

Nastanski, N. M., and Spell, T. L., 2004, Extracaldera Rhyolites North of the Yellowstone Plateau Volcanic Field Caldera Complex: An Evolving Silicic Magma System and Site of Future Large Volume Eruptions?: Abstracts with Programs, v. 36, no. 5, p. 431.

Spell, T. L., and Nastanski, N. M., 2004, Ion Microprobe ^^^Pb/^^U and ^^°Th/^^^U Zircon Ages From Extraealdera Rhyolites at Yellowstone: Constraints on Magma Residence Times and Evolution: Abstracts with programs, v. 36, no. 5, p. 431-432.

Druschke, P., Honn, D., McKelvey, M., Nastanski, N. M., Rager, A., Smith, E. I., Belliveau, R., 2004, Volcanology of the Northern Eldorado Mountains, Nevada: New Evidence for the Source of the Tufff of Bridge Spring?: Abstracts with programs, v. 36, no. 5, p. 431.

Nastanski, N. M^, and Spell, T. L., 2004, Do the Young Extraealdera Rhyolites North of Yellowstone Caldera Mark the Beginnings of a 4*’’ Volcanic Cycle in the Yellowston Plateau Volcanic Field?: Abstracts with programs, v. 36, no. 4, p. 11

Thesis Title: Petrogenesis of Extraealdera Rhyolites at Yellowstone Volcanic Field: Evidence for an Evolving Silicic Magma System North of Yellowstone Caldera

Thesis Examination Committee: Chairperson, Dr. Terry L. Spell, Ph.D. Committee Member, Dr. Eugene I. Smith, Ph.D. Committee Member, Dr. Rodney Metcalf, Ph.D. Graduate Faculty Representative, Dr. Steven DeBelle, Ph.D.

192

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.